@ai-on-browser/data-analysis-models 0.20.0

Default export object.

default

Properties

Tensor (Tensor) : Tensor class

Matrix (Matrix) : Matrix class

Graph (Graph) : Graph class

Complex (Complex) : Complex number

Static Members

▸ models

models

Properties

A2CAgent (A2CAgent) : A2C agent

LBABOD (LBABOD) : Lower-bound for the Angle-based Outlier Detection

ABOD (ABOD) : Angle-based Outlier Detection

ADALINE (ADALINE) : Adaptive Linear Neuron model

ADAMENN (ADAMENN) : Adaptive Metric Nearest Neighbor

AdaptiveThresholding (AdaptiveThresholding) : Adaptive thresholding

AffinityPropagation (AffinityPropagation) : Affinity propagation model

CentroidAgglomerativeClustering (CentroidAgglomerativeClustering) : Centroid agglomerative clustering

CompleteLinkageAgglomerativeClustering (CompleteLinkageAgglomerativeClustering) : Complete linkage agglomerative clustering

GroupAverageAgglomerativeClustering (GroupAverageAgglomerativeClustering) : Group average agglomerative clustering

MedianAgglomerativeClustering (MedianAgglomerativeClustering) : Median agglomerative clustering

SingleLinkageAgglomerativeClustering (SingleLinkageAgglomerativeClustering) : Single linkage agglomerative clustering

WardsAgglomerativeClustering (WardsAgglomerativeClustering) : Ward's agglomerative clustering

WeightedAverageAgglomerativeClustering (WeightedAverageAgglomerativeClustering) : Weighted average agglomerative clustering

AkimaInterpolation (AkimaInterpolation) : Akima interpolation

ALMA (ALMA) : Approximate Large Margin algorithm

AODE (AODE) : Averaged One-Dependence Estimators

AR (AR) : Autoregressive model

ARMA (ARMA) : Autoregressive moving average model

AROW (AROW) : Adaptive regularization of Weight Vectors

ART (ART) : Adaptive resonance theory

AssociationAnalysis (AssociationAnalysis) : Association analysis

Autoencoder (Autoencoder) : Autoencoder

AutomaticThresholding (AutomaticThresholding) : Automatic thresholding

AverageShiftedHistogram (AverageShiftedHistogram) : Average shifted histogram

BalancedHistogramThresholding (BalancedHistogramThresholding) : Balanced histogram thresholding

Ballseptron (Ballseptron) : Ballseptron

Banditron (Banditron) : Banditron

BayesianLinearRegression (BayesianLinearRegression) : Bayesian linear regression

BayesianNetwork (BayesianNetwork) : Bayesian Network

BernsenThresholding (BernsenThresholding) : Bernsen thresholding

BesselFilter (BesselFilter) : Bessel filter

BilinearInterpolation (BilinearInterpolation) : Bilinear interpolation

BIRCH (BIRCH) : Balanced iterative reducing and clustering using hierarchies

BisectingKMeans (BisectingKMeans) : Bisecting k-Means algorithm

BOGD (BOGD) : Bounded Online Gradient Descent

BoxCox (BoxCox) : Box-Cox transformation

BPA (BPA) : Budgeted online Passive-Aggressive

BrahmaguptaInterpolation (BrahmaguptaInterpolation) : Brahmagupta interpolation

BRIDGE (BRIDGE) : BRIDGE

MulticlassBSGD (MulticlassBSGD) : Multiclass Budgeted Stochastic Gradient Descent

BSGD (BSGD) : Budgeted Stochastic Gradient Descent

BudgetPerceptron (BudgetPerceptron) : Budget Perceptron

ButterworthFilter (ButterworthFilter) : Butterworth filter

C2P (C2P) : Clustering based on Closest Pairs

Canny (Canny) : Canny edge detection

CAST (CAST) : Clustering Affinity Search Technique

CategoricalNaiveBayes (CategoricalNaiveBayes) : Categorical naive bayes

CatmullRomSplines (CatmullRomSplines) : Catmull-Rom splines interpolation

CentripetalCatmullRomSplines (CentripetalCatmullRomSplines) : Centripetal Catmull-Rom splines interpolation

CHAMELEON (CHAMELEON) : CHAMELEON

ChangeFinder (ChangeFinder) : Change finder

ChebyshevFilter (ChebyshevFilter) : Chebyshev filter

CLARA (CLARA) : Clustering LARge Applications

CLARANS (CLARANS) : Clustering Large Applications based on RANdomized Search

CLIQUE (CLIQUE) : CLustering In QUEst

CLUES (CLUES) : CLUstEring based on local Shrinking

CoTraining (CoTraining) : Co-training

COF (COF) : Connectivity-based Outlier Factor

COLL (COLL) : Conscience on-line learning

ComplementNaiveBayes (ComplementNaiveBayes) : Complement Naive Bayes

ConfidenceWeighted (ConfidenceWeighted) : Confidence weighted

SoftConfidenceWeighted (SoftConfidenceWeighted) : Soft confidence weighted

CosineInterpolation (CosineInterpolation) : Cosine interpolation

CRF (CRF) : Conditional random fields

CubicConvolutionInterpolation (CubicConvolutionInterpolation) : Cubic-convolution interpolation

CubicHermiteSpline (CubicHermiteSpline) : Cubic Hermite spline

CubicInterpolation (CubicInterpolation) : Cubic interpolation

CumulativeMovingAverage (CumulativeMovingAverage) : Cumulative moving average

CumSum (CumSum) : Cumulative sum change point detection

CURE (CURE) : Clustering Using REpresentatives

DiscriminantAdaptiveNearestNeighbor (DiscriminantAdaptiveNearestNeighbor) : Discriminant adaptive nearest neighbor

DBCLASD (DBCLASD) : Distribution Based Clustering of LArge Spatial Databases

DBSCAN (DBSCAN) : Density-based spatial clustering of applications with noise

DecisionTreeClassifier (DecisionTreeClassifier) : Decision tree classifier

DecisionTreeRegression (DecisionTreeRegression) : Decision tree regression

DelaunayInterpolation (DelaunayInterpolation) : Delaunay interpolation

DemingRegression (DemingRegression) : Deming regression

DENCLUE (DENCLUE) : DENsity CLUstering

DIANA (DIANA) : DIvisive ANAlysis Clustering

DiffusionMap (DiffusionMap) : Diffusion map

DQNAgent (DQNAgent) : Deep Q-Network agent

DrakeKMeans (DrakeKMeans) : Drake's accelerated k-Means algorithm

DPAgent (DPAgent) : Dynamic programming agent

ElasticNet (ElasticNet) : Elastic net

ElkanKMeans (ElkanKMeans) : Elkan's accelerated k-Means algorithm

EllipticFilter (EllipticFilter) : Elliptic filter

ENaN (ENaN) : Extended Natural Neighbor

ENN (ENN) : Extended Nearest Neighbor

EnsembleBinaryModel (EnsembleBinaryModel) : Ensemble binary models

ExponentialMovingAverage (ExponentialMovingAverage) : Exponential moving average

ModifiedMovingAverage (ModifiedMovingAverage) : Modified moving average

ExtraTreesClassifier (ExtraTreesClassifier) : Extra trees classifier

ExtraTreesRegressor (ExtraTreesRegressor) : Extra trees regressor

FastMap (FastMap) : FastMap

FINDIT (FINDIT) : a Fast and INtelligent subspace clustering algorithm using DImension voting

Forgetron (Forgetron) : Forgetron

FuzzyCMeans (FuzzyCMeans) : Fuzzy c-means

FuzzyKNN (FuzzyKNN) : Fuzzy k-nearest neighbor

GAN (GAN) : Generative adversarial networks

GasserMuller (GasserMuller) : Gasser–Müller kernel estimator

GaussianProcess (GaussianProcess) : Gaussian process

GBDT (GBDT) : Gradient boosting decision tree

GBDTClassifier (GBDTClassifier) : Gradient boosting decision tree classifier

GeneralizedESD (GeneralizedESD) : Generalized extreme studentized deviate

GeneticAlgorithmGeneration (GeneticAlgorithmGeneration) : Genetic algorithm generation

GeneticKMeans (GeneticKMeans) : Genetic k-means model

GMeans (GMeans) : G-means

GMM (GMM) : Gaussian mixture model

GMR (GMR) : Gaussian mixture regression

SemiSupervisedGMM (SemiSupervisedGMM) : Semi-Supervised gaussian mixture model

GPLVM (GPLVM) : Gaussian Process Latent Variable Model

GrowingCellStructures (GrowingCellStructures) : Growing cell structures

GrowingNeuralGas (GrowingNeuralGas) : Growing neural gas

GSOM (GSOM) : Growing Self-Organizing Map

GTM (GTM) : Generative topographic mapping

HamelryKMeans (HamelryKMeans) : Hamelry's accelerated k-Means algorithm

HampelFilter (HampelFilter) : Hampel filter

HartiganWongKMeans (HartiganWongKMeans) : Hartigan-Wong k-Means algorithm

HDBSCAN (HDBSCAN) : Hierarchical Density-based spatial clustering of applications with noise

Histogram (Histogram) : Histogram

HLLE (HLLE) : Hessian Locally Linear Embedding

ContinuousHMM (ContinuousHMM) : Continuous hidden Markov model

HMM (HMM) : Hidden Markov model

HoltWinters (HoltWinters) : Holt-Winters method

HopfieldNetwork (HopfieldNetwork) : Hopfield network

Hotelling (Hotelling) : Hotelling T-square Method

HuberRegression (HuberRegression) : Huber regression

ICA (ICA) : Independent component analysis

CELLIP (CELLIP) : Classical ellipsoid method

IELLIP (IELLIP) : Improved ellipsoid method

IKNN (IKNN) : Locally Informative K-Nearest Neighbor

IncrementalPCA (IncrementalPCA) : Incremental principal component analysis

INFLO (INFLO) : Influenced Outlierness

InverseDistanceWeighting (InverseDistanceWeighting) : Inverse distance weighting

InverseSmoothstepInterpolation (InverseSmoothstepInterpolation) : Inverse smoothstep interpolation

ISODATA (ISODATA) : Iterative Self-Organizing Data Analysis Technique

IsolationForest (IsolationForest) : Isolation forest

Isomap (Isomap) : Isomap

IsotonicRegression (IsotonicRegression) : Isotonic regression

KalmanFilter (KalmanFilter) : Kalman filter

KDEOS (KDEOS) : Kernel Density Estimation Outlier Score

KernelDensityEstimator (KernelDensityEstimator) : Kernel density estimator

KernelKMeans (KernelKMeans) : Kernel k-means

KernelizedPegasos (KernelizedPegasos) : Kernelized Primal Estimated sub-GrAdientSOlver for SVM

KernelizedPerceptron (KernelizedPerceptron) : Kernelized perceptron

KHarmonicMeans (KHarmonicMeans) : K-Harmonic Means

KLIEP (KLIEP) : Kullback-Leibler importance estimation procedure

KMeans (KMeans) : k-means model

KMeanspp (KMeanspp) : k-means++ model

KMedians (KMedians) : k-medians model

KMedoids (KMedoids) : k-medoids model

SemiSupervisedKMeansModel (SemiSupervisedKMeansModel) : semi-supervised k-means model

KModes (KModes) : k-modes model

KNN (KNN) : k-nearest neighbor

KNNAnomaly (KNNAnomaly) : k-nearest neighbor anomaly detection

KNNDensityEstimation (KNNDensityEstimation) : k-nearest neighbor density estimation

KNNRegression (KNNRegression) : k-nearest neighbor regression

SemiSupervisedKNN (SemiSupervisedKNN) : Semi-supervised k-nearest neighbor

KPrototypes (KPrototypes) : k-prototypes model

KSVD (KSVD) : k-SVD

KolmogorovZurbenkoFilter (KolmogorovZurbenkoFilter) : Kolmogorov–Zurbenko filter

LabelPropagation (LabelPropagation) : Label propagation

LabelSpreading (LabelSpreading) : Label spreading

LadderNetwork (LadderNetwork) : Ladder network

LagrangeInterpolation (LagrangeInterpolation) : Lagrange interpolation

LanczosInterpolation (LanczosInterpolation) : Lanczos interpolation

Laplacian (Laplacian) : Laplacian edge detection

LaplacianEigenmaps (LaplacianEigenmaps) : Laplacian eigenmaps

Lasso (Lasso) : Least absolute shrinkage and selection operator

LatentDirichletAllocation (LatentDirichletAllocation) : Latent dirichlet allocation

LBG (LBG) : Linde-Buzo-Gray algorithm

FishersLinearDiscriminant (FishersLinearDiscriminant) : Fishers linear discriminant analysis

LinearDiscriminant (LinearDiscriminant) : Linear discriminant analysis

LinearDiscriminantAnalysis (LinearDiscriminantAnalysis) : Linear discriminant analysis

MulticlassLinearDiscriminant (MulticlassLinearDiscriminant) : Multiclass linear discriminant analysis

LDF (LDF) : Local Density Factor

LDOF (LDOF) : Local Distance-based Outlier Factor

LeastAbsolute (LeastAbsolute) : Least absolute deviations

LeastSquares (LeastSquares) : Least squares

LinearInterpolation (LinearInterpolation) : Linear interpolation

LLE (LLE) : Locally Linear Embedding

LeastMedianSquaresRegression (LeastMedianSquaresRegression) : Least median squares regression

LMNN (LMNN) : Large Margin Nearest Neighbor

LOCI (LOCI) : Local Correlation Integral

LOESS (LOESS) : Locally estimated scatterplot smoothing

LOF (LOF) : Local Outlier Factor

LoG (LoG) : Laplacian of gaussian filter

LogarithmicInterpolation (LogarithmicInterpolation) : Logarithmic interpolation

LogisticRegression (LogisticRegression) : Logistic regression

MultinomialLogisticRegression (MultinomialLogisticRegression) : Multinomial logistic regression

LoOP (LoOP) : Local Outlier Probability

LOWESS (LOWESS) : Locally weighted scatter plot smooth

LowpassFilter (LowpassFilter) : Lowpass filter

LpNormLinearRegression (LpNormLinearRegression) : Lp norm linear regression

LSA (LSA) : Latent Semantic Analysis

LSDD (LSDD) : Least-squares density difference

LSDDCPD (LSDDCPD) : LSDD for change point detection

LSIF (LSIF) : least-squares importance fitting

LeastTrimmedSquaresRegression (LeastTrimmedSquaresRegression) : Least trimmed squares

LTSA (LTSA) : Local Tangent Space Alignment

LVQClassifier (LVQClassifier) : Learning Vector Quantization classifier

LVQCluster (LVQCluster) : Learning Vector Quantization clustering

MacQueenKMeans (MacQueenKMeans) : MacQueen's k-Means algorithm

MAD (MAD) : Median Absolute Deviation

MADALINE (MADALINE) : Many Adaptive Linear Neuron model

MarginPerceptron (MarginPerceptron) : Margin Perceptron

MarkovSwitching (MarkovSwitching) : Markov switching

MaxAbsScaler (MaxAbsScaler) : Max absolute scaler

MaximumLikelihoodEstimator (MaximumLikelihoodEstimator) : Maximum likelihood estimator

MCD (MCD) : Minimum Covariance Determinant

MixtureDiscriminant (MixtureDiscriminant) : Mixture discriminant analysis

MDS (MDS) : Multi-dimensional Scaling

MeanShift (MeanShift) : Mean shift

MetropolisHastings (MetropolisHastings) : Metropolis-Hastings algorithm

MinmaxNormalization (MinmaxNormalization) : Min-max normalization

MIRA (MIRA) : Margin Infused Relaxed Algorithm

MLLE (MLLE) : Modified Locally Linear Embedding

MLPClassifier (MLPClassifier) : Multi layer perceptron classifier

MLPRegressor (MLPRegressor) : Multi layer perceptron regressor

MOD (MOD) : Method of Optimal Direction

MONA (MONA) : MONothetic Analysis Clustering

MonotheticClustering (MonotheticClustering) : Monothetic Clustering

MCAgent (MCAgent) : Monte Carlo agent

Mountain (Mountain) : Mountain method

LinearWeightedMovingAverage (LinearWeightedMovingAverage) : Linear weighted moving average

SimpleMovingAverage (SimpleMovingAverage) : Simple moving average

TriangularMovingAverage (TriangularMovingAverage) : Triangular moving average

MovingMedian (MovingMedian) : Moving median

MT (MT) : Mahalanobis Taguchi method

MutualInformationFeatureSelection (MutualInformationFeatureSelection) : Mutual information feature selector

MutualKNN (MutualKNN) : Mutual k-nearest-neighbor model

NCubicInterpolation (NCubicInterpolation) : n-cubic interpolation

NLinearInterpolation (NLinearInterpolation) : n-linear interpolation

NadarayaWatson (NadarayaWatson) : Nadaraya–Watson kernel regression

NaiveBayes (NaiveBayes) : Naive bayes

NaiveBayesRegression (NaiveBayesRegression) : Naive bayes regression

NAROW (NAROW) : Narrow Adaptive Regularization Of Weights

NaturalNeighborInterpolation (NaturalNeighborInterpolation) : Natural neighbor interpolation

NeighbourhoodComponentsAnalysis (NeighbourhoodComponentsAnalysis) : Neighbourhood components analysis

NearestCentroid (NearestCentroid) : Nearest centroid classifier

NegationNaiveBayes (NegationNaiveBayes) : Negation Naive bayes

NeuralGas (NeuralGas) : Neural gas model

ComputationalGraph (ComputationalGraph)

Layer (Layer)

NeuralnetworkException (NeuralnetworkException) : Exception for neuralnetwork class

NeuralNetwork (NeuralNetwork) : Neuralnetwork

NiblackThresholding (NiblackThresholding) : Niblack thresholding

NICE (NICE) : Flow-based generative model non-linear independent component estimation

NLMeans (NLMeans) : Non-local means filter

NMF (NMF) : Non-negative matrix factorization

NNBCA (NNBCA) : Natural Neighborhood Based Classification Algorithm

NOF (NOF) : Natural Outlier Factor

NormalHERD (NormalHERD) : Normal Herd

OAPBPM (OAPBPM) : Online Aggregate Prank-Bayes Point Machine

OCSVM (OCSVM) : One-class support vector machine

ODIN (ODIN) : Outlier Detection using Indegree Number

OnlineGradientDescent (OnlineGradientDescent) : Online gradient descent

OPTICS (OPTICS) : Ordering points to identify the clustering structure

ORCLUS (ORCLUS) : arbitrarily ORiented projected CLUSter generation

OrderedLogisticRegression (OrderedLogisticRegression) : Ordered logistic regression

OrderedProbitRegression (OrderedProbitRegression) : Ordered probit regression

OtsusThresholding (OtsusThresholding) : Otus's thresholding

PAM (PAM) : Partitioning Around Medoids

ParticleFilter (ParticleFilter) : Particle filter

PassingBablok (PassingBablok) : Passing-Bablok method

PA (PA) : Passive Aggressive

PAUM (PAUM) : Perceptron Algorithm with Uneven Margins

AnomalyPCA (AnomalyPCA) : Principal component analysis for anomaly detection

DualPCA (DualPCA) : Dual Principal component analysis

KernelPCA (KernelPCA) : Kernel Principal component analysis

PCA (PCA) : Principal component analysis

PossibilisticCMeans (PossibilisticCMeans) : Possibilistic c-means

PCR (PCR) : Principal component regression

Pegasos (Pegasos) : Primal Estimated sub-GrAdientSOlver for SVM

PercentileAnormaly (PercentileAnormaly) : Percentile anomaly detection

AveragedPerceptron (AveragedPerceptron) : Averaged perceptron

MulticlassPerceptron (MulticlassPerceptron) : Multiclass perceptron

Perceptron (Perceptron) : Perceptron

PhansalkarThresholding (PhansalkarThresholding) : Phansalkar thresholding

PLS (PLS) : Partial least squares regression

PLSA (PLSA) : Probabilistic latent semantic analysis

PoissonRegression (PoissonRegression) : Poisson regression

PGAgent (PGAgent) : Policy gradient agent

PolynomialHistogram (PolynomialHistogram) : Polynomial histogram

PolynomialInterpolation (PolynomialInterpolation) : Polynomial interpolation

ProjectionPursuit (ProjectionPursuit) : Projection pursuit regression

PRank (PRank) : Perceptron ranking

Prewitt (Prewitt) : Prewitt edge detection

PriestleyChao (PriestleyChao) : Priestley–Chao kernel estimator

PrincipalCurve (PrincipalCurve) : Principal curves

ProbabilisticPCA (ProbabilisticPCA) : Probabilistic Principal component analysis

ProbabilityBasedClassifier (ProbabilityBasedClassifier) : Probability based classifier

MultinomialProbit (MultinomialProbit) : Multinomial probit

Probit (Probit) : Probit

PROCLUS (PROCLUS) : PROjected CLUStering algorithm

Projectron (Projectron) : Projectron

Projectronpp (Projectronpp) : Projectron++

PTile (PTile) : P-tile thresholding

QTableBase (QTableBase) : Base class for Q-table

QAgent (QAgent) : Q-learning agent

QuadraticDiscriminant (QuadraticDiscriminant) : Quadratic discriminant analysis

QuantileRegression (QuantileRegression) : Quantile regression

RadiusNeighbor (RadiusNeighbor) : radius neighbor

RadiusNeighborRegression (RadiusNeighborRegression) : radius neighbor regression

SemiSupervisedRadiusNeighbor (SemiSupervisedRadiusNeighbor) : Semi-supervised radius neighbor

RamerDouglasPeucker (RamerDouglasPeucker) : Ramer-Douglas-Peucker algorithm

RandomForestClassifier (RandomForestClassifier) : Random forest classifier

RandomForestRegressor (RandomForestRegressor) : Random forest regressor

RandomProjection (RandomProjection) : Random projection

RankNet (RankNet) : RankNet

RANSAC (RANSAC) : Random sample consensus

RadialBasisFunctionNetwork (RadialBasisFunctionNetwork) : Radial basis function network

GBRBM (GBRBM) : Gaussian-Bernouili Restricted Boltzmann machine

RBM (RBM) : Restricted Boltzmann machine

RBP (RBP) : Randomized Budget Perceptron

RDF (RDF) : Relative Density Factor

RDOS (RDOS) : Relative Density-based Outlier Score

KernelRidge (KernelRidge) : Kernel ridge regression

MulticlassRidge (MulticlassRidge) : Multiclass ridge regressioin

Ridge (Ridge) : Ridge regressioin

RKOF (RKOF) : Robust Kernel-based Outlier Factor

RecursiveLeastSquares (RecursiveLeastSquares) : Recursive least squares

RepeatedMedianRegression (RepeatedMedianRegression) : Repeated median regression

RNN (RNN) : Recurrent neuralnetwork

RobertsCross (RobertsCross) : Roberts cross

RobustScaler (RobustScaler) : Robust scaler

ROCK (ROCK) : RObust Clustering using linKs

AggressiveROMMA (AggressiveROMMA) : Aggressive Relaxed Online Maximum Margin Algorithm

ROMMA (ROMMA) : Relaxed Online Maximum Margin Algorithm

RVM (RVM) : Relevance vector machine

S3VM (S3VM) : Semi-Supervised Support Vector Machine

Sammon (Sammon) : Sammon mapping

SARSAAgent (SARSAAgent) : SARSA agent

SauvolaThresholding (SauvolaThresholding) : sauvola thresholding

SavitzkyGolayFilter (SavitzkyGolayFilter) : Savitzky-Golay filter

SDAR (SDAR) : Sequentially Discounting Autoregressive model

SegmentedRegression (SegmentedRegression) : Segmented regression

SelectiveNaiveBayes (SelectiveNaiveBayes) : Selective Naive bayes

SelectiveSamplingAdaptivePerceptron (SelectiveSamplingAdaptivePerceptron) : Selective sampling Perceptron with adaptive parameter

SelectiveSamplingPerceptron (SelectiveSamplingPerceptron) : Selective sampling Perceptron

SelectiveSamplingSOP (SelectiveSamplingSOP) : Selective sampling second-order Perceptron

SelectiveSamplingWinnow (SelectiveSamplingWinnow) : Selective sampling Winnow

SelfTraining (SelfTraining) : Self-training

SemiSupervisedNaiveBayes (SemiSupervisedNaiveBayes) : Semi-supervised naive bayes

SezanThresholding (SezanThresholding) : Sezan's thresholding

ShiftingPerceptron (ShiftingPerceptron) : Shifting Perceptron Algorithm

ILK (ILK) : Implicit online Learning with Kernels

SILK (SILK) : Sparse Implicit online Learning with Kernels

SincInterpolation (SincInterpolation) : Sinc interpolation

SlicedInverseRegression (SlicedInverseRegression) : Sliced inverse regression

Slerp (Slerp) : Spherical linear interpolation

SliceSampling (SliceSampling) : slice sampling

SMARegression (SMARegression) : Standardizes Major Axis regression

SmirnovGrubbs (SmirnovGrubbs) : SmirnovGrubbs test

SmoothstepInterpolation (SmoothstepInterpolation) : Smoothstep interpolation

Snakes (Snakes) : Snakes (active contour model)

Sobel (Sobel) : Sobel edge detection

SoftKMeans (SoftKMeans) : Soft k-means

SOM (SOM) : Self-Organizing Map

SecondOrderPerceptron (SecondOrderPerceptron) : Second order perceptron

SpectralClustering (SpectralClustering) : Spectral clustering

SmoothingSpline (SmoothingSpline) : Spline smoothing

SplineInterpolation (SplineInterpolation) : Spline interpolation

SplitAndMerge (SplitAndMerge) : Split and merge segmentation

SquaredLossMICPD (SquaredLossMICPD) : Squared-loss Mutual information change point detection

SST (SST) : Singular-spectrum transformation

Standardization (Standardization) : Standardization

StatisticalRegionMerging (StatisticalRegionMerging) : Statistical Region Merging

STING (STING) : STatistical INformation Grid-based method

Stoptron (Stoptron) : Stoptron

SVC (SVC) : Support vector clustering

SVM (SVM) : Support vector machine

SVR (SVR) : Support vector regression

TheilSenRegression (TheilSenRegression) : Theil-Sen regression

Thompson (Thompson) : Thompson test

TietjenMoore (TietjenMoore) : Tietjen-Moore Test

TighterPerceptron (TighterPerceptron) : Tighter Budget Perceptron

TightestPerceptron (TightestPerceptron) : Tightest Perceptron

TrigonometricInterpolation (TrigonometricInterpolation) : Trigonometric interpolation

SNE (SNE) : Stochastic Neighbor Embedding

tSNE (tSNE) : T-distributed Stochastic Neighbor Embedding

TukeyRegression (TukeyRegression) : Tukey regression

TukeysFences (TukeysFences) : Tukey's fences

RuLSIF (RuLSIF) : Relative unconstrained Least-Squares Importance Fitting

uLSIF (uLSIF) : unconstrained Least-Squares Importance Fitting

UMAP (UMAP) : Uniform Manifold Approximation and Projection

UniversalSetNaiveBayes (UniversalSetNaiveBayes) : Universal-set Naive bayes

VAE (VAE) : Variational Autoencoder

VAR (VAR) : Vector Autoregressive model

VBGMM (VBGMM) : Variational Gaussian Mixture Model

VotedPerceptron (VotedPerceptron) : Voted-perceptron

WeightedKMeans (WeightedKMeans) : Weighted k-means model

WeightedKNN (WeightedKNN) : Weighted K-Nearest Neighbor

WeightedLeastSquares (WeightedLeastSquares) : Weighted least squares

Winnow (Winnow) : Winnow

Word2Vec (Word2Vec) : Word2Vec

XGBoost (XGBoost) : eXtreme Gradient Boosting regression

XGBoostClassifier (XGBoostClassifier) : eXtreme Gradient Boosting classifier

XMeans (XMeans) : x-means

YeoJohnson (YeoJohnson) : Yeo-Johnson power transformation

YinyangKMeans (YinyangKMeans) : Yinyang k-Means algorithm

ZeroInflatedNegativeBinomial (ZeroInflatedNegativeBinomial) : Zero-inflated negative binomial

ZeroInflatedPoisson (ZeroInflatedPoisson) : Zero-inflated poisson

ZeroTruncatedPoisson (ZeroTruncatedPoisson) : Zero-truncated poisson

▸ rl

rl

Properties

AcrobotRLEnvironment (AcrobotRLEnvironment) : Acrobot environment

RLEnvironmentBase (RLEnvironmentBase) : Base class for reinforcement learning environment

RLIntRange (RLIntRange) : Integer number range state/actioin

RLRealRange (RLRealRange) : Real number range state/actioin

EmptyRLEnvironment (EmptyRLEnvironment) : Empty environment

BlackjackRLEnvironment (BlackjackRLEnvironment) : Blackjack environment

BreakerRLEnvironment (BreakerRLEnvironment) : Breaker environment

CartPoleRLEnvironment (CartPoleRLEnvironment) : Cartpole environment

DraughtsRLEnvironment (DraughtsRLEnvironment) : Draughts environment

GomokuRLEnvironment (GomokuRLEnvironment) : Gomoku environment

GridMazeRLEnvironment (GridMazeRLEnvironment) : Grid world environment

InHypercubeRLEnvironment (InHypercubeRLEnvironment) : In-hypercube environment

SmoothMazeRLEnvironment (SmoothMazeRLEnvironment) : Smooth maze environment

MountainCarRLEnvironment (MountainCarRLEnvironment) : MountainCar environment

PendulumRLEnvironment (PendulumRLEnvironment) : Pendulum environment

ReversiRLEnvironment (ReversiRLEnvironment) : Reversi environment

WaterballRLEnvironment (WaterballRLEnvironment) : Waterball environment

▸ evaluate

evaluate

Properties

accuracy (accuracy) : Returns accuracy.

cohensKappa (cohensKappa) : Returns Cohen's kappa coefficient.

fScore (fScore) : Returns F-score with macro average.

precision (precision) : Returns precision with macro average.

recall (recall) : Returns recall with macro average.

davisBouldinIndex (davisBouldinIndex) : Returns Davies-Bouldin index.

diceIndex (diceIndex) : Returns Dice index.

dunnIndex (dunnIndex) : Returns Dunn index.

fowlkesMallowsIndex (fowlkesMallowsIndex) : Returns Fowlkes-Mallows index.

jaccardIndex (jaccardIndex) : Returns Jaccard index.

purity (purity) : Returns Purity.

randIndex (randIndex) : Returns Rand index.

silhouetteCoefficient (silhouetteCoefficient) : Returns Silhouette coefficient.

coRankingMatrix (coRankingMatrix) : Returns Co-Ranking Matrix.

correlation (correlation) : Returns correlation.

mad (mad) : Returns MAD (Median Absolute Deviation).

mae (mae) : Returns MAE (Mean Absolute Error).

mape (mape) : Returns MAPE (Mean Absolute Percentage Error).

mse (mse) : Returns MSE (Mean Squared Error).

msle (msle) : Returns MSLE (Mean Squared Logarithmic Error).

r2 (r2) : Returns R2 (coefficient of determination).

rmse (rmse) : Returns RMSE (Root Mean Squared Error).

rmsle (rmsle) : Returns RMSLE (Root Mean Squared Logarithmic Error).

rmspe (rmspe) : Returns RMSPE (Root Mean Squared Percentage Error).

MatrixException

Exception for matrix class

new MatrixException(message: string, value: any)

Extends Error

Parameters

message (string) Error message

value (any) Some value

Matrix

Matrix class

new Matrix(rows: number, cols: number, values: (number | Array<number> | Array<Array<number>>)?)

Parameters

rows (number) Number of rows

cols (number) Number of columns

values ((number | Array<number> | Array<Array<number>>)?) Initial values

Static Members

▸ zeros(rows, cols)

Returns a matrix filled with 0.

zeros(rows: number, cols: number): Matrix

Parameters

rows (number) Number of rows

cols (number) Number of columns

Returns

Matrix: Matrix filled with 0

▸ ones(rows, cols)

Returns a matrix filled with 1.

ones(rows: number, cols: number): Matrix

Parameters

rows (number) Number of rows

cols (number) Number of columns

Returns

Matrix: Matrix filled with 1

▸ eye(rows, cols, init = 1)

Returns a identity matrix.

eye(rows: number, cols: number, init: number?): Matrix

Parameters

rows (number) Number of rows

cols (number) Number of columns

init (number? = 1) Diagonal values

Returns

Matrix: Identity matrix

▸ random(rows, cols, min = 0, max = 1)

Returns a matrix initialized uniform random values.

random(rows: number, cols: number, min: number?, max: number?): Matrix

Parameters

rows (number) Number of rows

cols (number) Number of columns

min (number? = 0) Minimum value of the Matrix (include)

max (number? = 1) Maximum value of the Matrix (exclude)

Returns

Matrix: Matrix initialized uniform random values

▸ randint(rows, cols, min = 0, max = 1)

Returns a matrix initialized uniform random integer values.

randint(rows: number, cols: number, min: number?, max: number?): Matrix

Parameters

rows (number) Number of rows

cols (number) Number of columns

min (number? = 0) Minimum value of the Matrix (include)

max (number? = 1) Maximum value of the Matrix (include)

Returns

Matrix: Matrix initialized uniform random values

▸ randn(rows, cols, myu = 0, sigma = 1)

Returns a matrix initialized normal random values.

randn(rows: number, cols: number, myu: (number | Array<number>)?, sigma: (number | Array<Array<number>>)?): Matrix

Parameters

rows (number) Number of rows

cols (number) Number of columns

myu ((number | Array<number>)? = 0) Mean value(s) of each columns

sigma ((number | Array<Array<number>>)? = 1) Variance value or covariance matrix of each columns

Returns

Matrix: Matrix initialized normal random values

▸ diag(d)

Returns a diagonal matrix.

diag(d: Array<(number | Matrix)>): Matrix

Parameters

d (Array<(number | Matrix)>) Diagonal values

Returns

Matrix: Diagonal matrix

▸ fromArray(arr)

Returns a matrix from some value.

fromArray(arr: (Matrix | Array<Array<number>> | Array<number> | number)): Matrix

Parameters

arr ((Matrix | Array<Array<number>> | Array<number> | number)) Original values

Returns

Matrix: Matrix from some value

▸ map(mat, cb)

Returns a matrix that replace all the elements.

map(mat: Matrix, cb: function (number, Array<number>, Matrix): number): Matrix

Parameters

mat (Matrix) Original matrix

cb (function (number, Array<number>, Matrix): number) Mapping function

Returns

Matrix: Mapped matrix

▸ resize(mat, rows, cols, init = 0)

Return resized matrix.

resize(mat: Matrix, rows: number, cols: number, init: number?): Matrix

Parameters

mat (Matrix) Original matrix

rows (number) New row size

cols (number) New column size

init (number? = 0) Value of the extended region

Returns

Matrix: Resized matrix

▸ repeat(mat, n, axis = 0)

Returns a matrix that repeat the elements n times along the axis.

repeat(mat: Matrix, n: (number | Array<number>), axis: number?): Matrix

Parameters

mat (Matrix) Original matrix

n ((number | Array<number>)) Repeated count(s)

axis (number? = 0) Axis to be repeated

Returns

Matrix: Repeated matrix

▸ concat(a, b, axis = 0)

Returns a matrix concatenated this and m.

concat(a: Matrix, b: Matrix, axis: number?): Matrix

Parameters

a (Matrix) Original matrix

b (Matrix) Concatenate matrix

axis (number? = 0) Axis to be concatenated

Returns

Matrix: Concatenated matrix

▸ add(a, b)

Returns a matrix that add two values.

add(a: (Matrix | number), b: (Matrix | number)): Matrix

Parameters

a ((Matrix | number)) Left value

b ((Matrix | number)) Right value

Returns

Matrix: Added matrix

▸ sub(a, b)

Returns a matrix that subtract two values.

sub(a: (Matrix | number), b: (Matrix | number)): Matrix

Parameters

a ((Matrix | number)) Left value

b ((Matrix | number)) Right value

Returns

Matrix: Subtracted matrix

▸ mult(a, b)

Returns a matrix that multiplies by two values element-wise.

mult(a: (Matrix | number), b: (Matrix | number)): Matrix

Parameters

a ((Matrix | number)) Left value

b ((Matrix | number)) Right value

Returns

Matrix: Multiplied matrix

▸ div(a, b)

Returns a matrix that divides by two values element-wise.

div(a: (Matrix | number), b: (Matrix | number)): Matrix

Parameters

a ((Matrix | number)) Left value

b ((Matrix | number)) Right value

Returns

Matrix: Divided matrix

▸ mod(a, b)

Returns a matrix that takes a remainder divided by two values element-wise.

mod(a: (Matrix | number), b: (Matrix | number)): Matrix

Parameters

a ((Matrix | number)) Left value

b ((Matrix | number)) Right value

Returns

Matrix: Remainder matrix

▸ and(a, b)

Returns a matrix that takes logical AND two values.

and(a: (Matrix | number), b: (Matrix | number)): Matrix

Parameters

a ((Matrix | number)) Left value

b ((Matrix | number)) Right value

Returns

Matrix: Logical AND matrix

▸ or(a, b)

Returns a matrix that takes logical OR two values.

or(a: (Matrix | number), b: (Matrix | number)): Matrix

Parameters

a ((Matrix | number)) Left value

b ((Matrix | number)) Right value

Returns

Matrix: Logical OR matrix

▸ bitand(a, b)

Returns a matrix that takes bitwise AND two values.

bitand(a: (Matrix | number), b: (Matrix | number)): Matrix

Parameters

a ((Matrix | number)) Left value

b ((Matrix | number)) Right value

Returns

Matrix: Bitwise AND matrix

▸ bitor(a, b)

Returns a matrix that takes bitwise OR two values.

bitor(a: (Matrix | number), b: (Matrix | number)): Matrix

Parameters

a ((Matrix | number)) Left value

b ((Matrix | number)) Right value

Returns

Matrix: Bitwise OR matrix

▸ bitxor(a, b)

Returns a matrix that takes bitwise XOR two values.

bitxor(a: (Matrix | number), b: (Matrix | number)): Matrix

Parameters

a ((Matrix | number)) Left value

b ((Matrix | number)) Right value

Returns

Matrix: Bitwise XOR matrix

Instance Members

▸ dimension

Dimension of the matrix.

dimension

Type: number

▸ sizes

Sizes of the matrix.

sizes

Type: Array<number>

▸ length

Number of all elements in the matrix.

length

Type: number

▸ rows

Number of rows of the matrix.

rows

Type: number

▸ cols

Number of columns of the matrix.

cols

Type: number

▸ value

Elements in the matrix.

value

Type: Array<number>

▸ t

Transpose matrix.

t

Type: Matrix

▸ iterator()

Iterate over the elements.

iterator()

▸ toArray()

Returns a nested array represented this matrix.

toArray(): Array<Array<number>>

Returns

Array<Array<number>>: Nested array

▸ toScaler()

Returns the only element.

toScaler(): number

Returns

number: The only element

▸ toString()

Returns a string represented this matrix.

toString(): string

Returns

string: String represented this matrix

▸ copy(dst?)

Returns a copy of this matrix.

copy(dst: Matrix?): Matrix

Parameters

dst (Matrix?) Destination matrix

Returns

Matrix: Copied matrix

▸ equals(other, tol = 0)

Returns this matrix is equals to the others.

equals(other: any, tol: number?): boolean

Parameters

other (any) Check tensor

tol (number? = 0) Tolerance to be recognized as the same

Returns

boolean: true if equal

▸ at(r, c?)

Returns a value at the position.

at(r: (number | [number, number]), c: number?): number

Parameters

r ((number | [number, number])) Row index or index values

c (number?) Column index

Returns

number: Value at the position

▸ set(r, c, value?)

Set a value at the position.

set(r: (number | [number, number]), c: (number | Matrix), value: (number | Matrix)?)

Parameters

r ((number | [number, number])) Row index or index values. If this value is an array, the next argument should be the value to be set

c ((number | Matrix)) Column index or the value to be set

value ((number | Matrix)?) The value to be set

▸ row(r)

Returns a row matrix at r.

row(r: (number | Array<number> | Array<boolean>)): Matrix

Parameters

r ((number | Array<number> | Array<boolean>)) Indexes of rows, or an array of boolean values where the row to be selected is true.

Returns

Matrix: Row selected matrix

▸ col(c)

Returns a col matrix at c.

col(c: (number | Array<number> | Array<boolean>)): Matrix

Parameters

c ((number | Array<number> | Array<boolean>)) Indexes of columns, or an array of boolean values where the column to be selected is true.

Returns

Matrix: Column selected matrix

▸ slice(from, to, axis = 0)

Returns sliced matrix.

slice(from: number, to: number, axis: number?): Matrix

Parameters

from (number) Start index

to (number) End index

axis (number? = 0) Axis to be sliced

Returns

Matrix: Sliced matrix

▸ block(rows_from?, cols_from?, rows_to?, cols_to?)

Returns the sub-matrix corresponding to position.

block(rows_from: number?, cols_from: number?, rows_to: number?, cols_to: number?): Matrix

Parameters

rows_from (number?) Start row index

cols_from (number?) Start column index

rows_to (number?) End row index(exclusive)

cols_to (number?) End column index(exclusive)

Returns

Matrix: Sub matrix

▸ remove(idx, axis = 0)

Remove specified indexes.

remove(idx: (number | Array<number>), axis: number?)

Parameters

idx ((number | Array<number>)) Remove index

axis (number? = 0) Axis to be removed

▸ removeIf(cond, axis = 0)

Remove specified indexes.

removeIf(cond: function (Matrix): boolean, axis: number?)

Parameters

cond (function (Matrix): boolean) Remove condition function. Remove if it returns true

axis (number? = 0) Axis to be removed

▸ sample(n, axis = 0)

Returns a matrix that sampled along the axis.

sample(n: number, axis: number?): [Matrix, Array<number>]

Parameters

n (number) Sampled size

axis (number? = 0) Axis to be sampled

Returns

[Matrix, Array<number>]: Sampled matrix and its original indexes

▸ fill(value)

Fill in all the elements with the value.

fill(value: number)

Parameters

value (number) Filled value

▸ map(cb)

Iterate over all the elements and replace the value.

map(cb: function (number, Array<number>, Matrix): number)

Parameters

cb (function (number, Array<number>, Matrix): number) Mapping function

▸ forEach(cb)

Iterate over all the elements.

forEach(cb: function (number, Array<number>, Matrix): number)

Parameters

cb (function (number, Array<number>, Matrix): number) Callback function

▸ transpose()

Returns transpose matrix.

transpose(): Matrix

Returns

Matrix: Transposed matrix

▸ adjoint()

Returns adjoint matrix.

adjoint(): Matrix

Returns

Matrix: Adjoint matrix

▸ flip(axis = 0)

Flip values along the axis.

flip(axis: number?)

Parameters

axis (number? = 0) Axis to be flipped

▸ swap(a, b, axis = 0)

Swap the index a and b along the axis.

swap(a: number, b: number, axis: number?)

Parameters

a (number) First index

b (number) Second index

axis (number? = 0) Axis to be swapped

▸ sort(axis = 0)

Sort values along the axis.

sort(axis: number?): Array<number>

Parameters

axis (number? = 0) Axis to be sorted

Returns

Array<number>: Original index.

▸ shuffle(axis = 0)

Shuffle values along the axis.

shuffle(axis: number?): Array<number>

Parameters

axis (number? = 0) Axis to be shuffled

Returns

Array<number>: Original index.

▸ unique(axis = 0, tol = 0)

Make it unique in the specified axis.

unique(axis: number?, tol: number?): Array<number>

Parameters

axis (number? = 0) Axis to be uniqued

tol (number? = 0) Tolerance to be recognized as the same

Returns

Array<number>: Selected indexes

▸ resize(rows, cols, init = 0)

Resize this matrix.

resize(rows: number, cols: number, init: number?)

Parameters

rows (number) New row size

cols (number) New column size

init (number? = 0) Value of the extended region

▸ reshape(rows, cols)

Reshape this.

reshape(rows: number, cols: number)

Parameters

rows (number) New row size

cols (number) New column size

▸ repeat(n, axis = 0)

Repeat the elements n times along the axis this.

repeat(n: (number | Array<number>), axis: number?)

Parameters

n ((number | Array<number>)) Repeated count(s)

axis (number? = 0) Axis to be repeated

▸ concat(m, axis = 0)

Concatenate this and m.

concat(m: Matrix, axis: number?)

Parameters

m (Matrix) Concatenate matrix

axis (number? = 0) Axis to be concatenated

▸ reduce(cb, init?, axis = -1, keepdims = null)

Returns a matrix reduced along the axis with the callback function.

reduce(cb: function (number, number, Array<number>, Matrix): number, init: any?, axis: (number | Array<number>)?, keepdims: boolean?): (Matrix | number)

Parameters

cb (function (number, number, Array<number>, Matrix): number) Reducing function

init (any?) Initial value

axis ((number | Array<number>)? = -1) Axis to be reduced. If negative, reduce along all elements.

keepdims (boolean? = null) Keep dimensions or not. If null, negative axis retuns number and other axis returns Matrix.

Returns

(Matrix | number): Reduced matrix or value

▸ every(cb, axis = -1)

Determines whether all the members of a matrix satisfy the specified test.

every(cb: function (number, Array<number>, Matrix): number, axis: number?): (boolean | Matrix)

Parameters

cb (function (number, Array<number>, Matrix): number) Check function

axis (number? = -1) Axis to be reduced

Returns

(boolean | Matrix): Reduced value or matrix

▸ some(cb, axis = -1)

Determines whether the specified callback function returns true for any element of a matrix.

some(cb: function (number, Array<number>, Matrix): number, axis: number?): (boolean | Matrix)

Parameters

cb (function (number, Array<number>, Matrix): number) Check function

axis (number? = -1) Axis to be reduced

Returns

(boolean | Matrix): Reduced value or matrix

▸ max(axis = -1)

Returns maximum values along the axis.

max(axis: number?): (Matrix | number)

Parameters

axis (number? = -1) Axis to be reduced. If negative, returns the maximum value of the all element.

Returns

(Matrix | number): Maximum values

▸ min(axis = -1)

Returns minimum values along the axis.

min(axis: number?): (Matrix | number)

Parameters

axis (number? = -1) Axis to be reduced. If negative, returns the minimum value of the all element.

Returns

(Matrix | number): Minimum values

▸ median(axis = -1)

Returns medians along the axis.

median(axis: number?): (Matrix | number)

Parameters

axis (number? = -1) Axis to be reduced. If negative, returns a median of the all element.

Returns

(Matrix | number): Median values

▸ quantile(q, axis = -1)

Returns quantile values along the axis.

quantile(q: number, axis: number?): (Matrix | number)

Parameters

q (number) Partition rate

axis (number? = -1) Axis to be reduced. If negative, returns the quantile value of the all element.

Returns

(Matrix | number): Quantile values

▸ argmax(axis)

Returns maximum indexes along the axis.

argmax(axis: number): Matrix

Parameters

axis (number) Axis to be reduced

Returns

Matrix: Argmax values

▸ argmin(axis)

Returns minimum indexes along the axis.

argmin(axis: number): Matrix

Parameters

axis (number) Axis to be reduced

Returns

Matrix: Argmin values

▸ sum(axis = -1)

Returns summation values along the axis.

sum(axis: number?): (Matrix | number)

Parameters

axis (number? = -1) Axis to be reduced. If negative, returns a summation value of the all element.

Returns

(Matrix | number): Summation values

▸ mean(axis = -1)

Returns means along the axis.

mean(axis: number?): (Matrix | number)

Parameters

axis (number? = -1) Axis to be reduced. If negative, returns a mean value of the all element.

Returns

(Matrix | number): Mean values

▸ prod(axis = -1)

Returns producted values along the axis.

prod(axis: number?): (Matrix | number)

Parameters

axis (number? = -1) Axis to be reduced. If negative, returns a producted value of the all element.

Returns

(Matrix | number): Producted values

▸ variance(axis = -1, ddof = 0)

Returns variances along the axis.

variance(axis: number?, ddof: number?): (Matrix | number)

Parameters

axis (number? = -1) Axis to be reduced. If negative, returns a variance of the all element.

ddof (number? = 0) Delta Degrees of Freedom

Returns

(Matrix | number): Variance values

▸ std(axis = -1, ddof = 0)

Returns standard deviations along the axis.

std(axis: number?, ddof: number?): (Matrix | number)

Parameters

axis (number? = -1) Axis to be reduced. If negative, returns a standard deviation of the all element.

ddof (number? = 0) Delta Degrees of Freedom

Returns

(Matrix | number): Standard deviation values

▸ isSquare()

Returns if this is square matrix or not.

isSquare(): boolean

Returns

boolean: true if this is square matrix

▸ isDiag(tol = 0)

Returns if this is diagonal matrix or not.

isDiag(tol: number?): boolean

Parameters

tol (number? = 0) Tolerance to be recognized as 0

Returns

boolean: true if this is diagonal matrix

▸ isIdentity(tol = 0)

Returns if this is identity matrix or not.

isIdentity(tol: number?): boolean

Parameters

tol (number? = 0) Tolerance to be recognized as 0 or 1

Returns

boolean: true if this is identity matrix

▸ isZero(tol = 0)

Returns if this is zero matrix or not.

isZero(tol: number?): boolean

Parameters

tol (number? = 0) Tolerance to be recognized as 0

Returns

boolean: true if this is zero matrix

▸ isTriangular(tol = 0)

Returns if this is triangular matrix or not.

isTriangular(tol: number?): boolean

Parameters

tol (number? = 0) Tolerance to be recognized as 0

Returns

boolean: true if this is triangular matrix

▸ isLowerTriangular(tol = 0)

Returns if this is lower triangular matrix or not.

isLowerTriangular(tol: number?): boolean

Parameters

tol (number? = 0) Tolerance to be recognized as 0

Returns

boolean: true if this is lower triangular matrix

▸ isUpperTriangular(tol = 0)

Returns if this is upper triangular matrix or not.

isUpperTriangular(tol: number?): boolean

Parameters

tol (number? = 0) Tolerance to be recognized as 0

Returns

boolean: true if this is upper triangular matrix

▸ isSymmetric(tol = 0)

Returns if this is symmetric matrix or not.

isSymmetric(tol: number?): boolean

Parameters

tol (number? = 0) Tolerance to be recognized as the same

Returns

boolean: true if this is symmetric matrix

▸ isHermitian(tol = 0)

Returns if this is hermitian matrix or not.

isHermitian(tol: number?): boolean

Parameters

tol (number? = 0) Tolerance to be recognized as the same

Returns

boolean: true if this is hermitian matrix

▸ isAlternating(tol = 0)

Returns if this is alternating matrix or not.

isAlternating(tol: number?): boolean

Parameters

tol (number? = 0) Tolerance within which sign-reversed values are recognized as the same

Returns

boolean: true if this is alternating matrix

▸ isSkewHermitian(tol = 0)

Returns if this is skew-hermitian matrix or not.

isSkewHermitian(tol: number?): boolean

Parameters

tol (number? = 0) Tolerance within which sign-reversed values are recognized as the same

Returns

boolean: true if this is skew-hermitian matrix

▸ isRegular(tol = 0)

Returns if this is regular matrix or not.

isRegular(tol: number?): boolean

Parameters

tol (number? = 0) Tolerance to recognize the determinant as 0

Returns

boolean: true if this is regular matrix

▸ isNormal(tol = 0)

Returns if this is normal matrix or not.

isNormal(tol: number?): boolean

Parameters

tol (number? = 0) Tolerance to be recognized as the same

Returns

boolean: true if this is normal matrix

▸ isOrthogonal(tol = 0)

Returns if this is orthogonal matrix or not.

isOrthogonal(tol: number?): boolean

Parameters

tol (number? = 0) Tolerance to be recognized as 0 or 1

Returns

boolean: true if this is orthogonal matrix

▸ isUnitary(tol = 0)

Returns if this is unitary matrix or not.

isUnitary(tol: number?): boolean

Parameters

tol (number? = 0) Tolerance to be recognized as 0 or 1

Returns

boolean: true if this is unitary matrix

▸ isNilpotent(tol = 0)

Returns if this is nilpotent matrix or not.

isNilpotent(tol: number?): boolean

Parameters

tol (number? = 0) Tolerance to be recognized as 0

Returns

boolean: true if this is nilpotent matrix

▸ diag()

Returns diagonal elements.

diag(): Array<number>

Returns

Array<number>: Diagonal values

▸ trace()

Returns a trace.

trace(): number

Returns

number: Trace value

▸ norm(p = 2)

Returns a p-norm.

norm(p: number?): number

Parameters

p (number? = 2) p-norm

Returns

number: Entry-wise norm

▸ normInduced(p = 2)

Returns induced norm.

normInduced(p: number?): number

Parameters

p (number? = 2) p-norm

Returns

number: Induced norm

▸ normSpectral()

Returns spectral norm.

normSpectral(): number

Returns

number: Spectral norm

▸ normEntrywise(p = 2)

Returns a entry-wise norm

normEntrywise(p: number?): number

Parameters

p (number? = 2) p-norm

Returns

number: Entry-wise norm

▸ normFrobenius()

Returns frobenius norm.

normFrobenius(): number

Returns

number: Frobenius norm

▸ normMax()

Returns max norm.

normMax(): number

Returns

number: Max norm

▸ normSchatten(p = 2)

Returns schatten norm.

normSchatten(p: number?): number

Parameters

p (number? = 2) p-norm

Returns

number: Schatten norm

▸ normNuclear()

Returns nuclear norm.

normNuclear(): number

Returns

number: Nuclear norm

▸ rank(tol = 0)

Returns a rank of this matrix.

rank(tol: number?): number

Parameters

tol (number? = 0) Tolerance to be recognized as the same

Returns

number: Rank of this matrix

▸ det()

Returns a determinant.

det(): number

Returns

number: Determinant value

▸ negative()

Multiply all elements by -1 in-place.

negative()

▸ not()

Set all elements to their logical NOT values.

not()

▸ bitnot()

Set all elements to their bitwise NOT values.

bitnot()

▸ abs()

Set all elements to their absolute values.

abs()

▸ round()

Set all elements to their rounded values.

round()

▸ floor()

Set all elements to their floored values.

floor()

▸ ceil()

Set all elements to their ceil values.

ceil()

▸ leftShift(n)

Set all elements to their left shift values.

leftShift(n: number)

Parameters

n (number) Shift amount

▸ signedRightShift(n)

Set all elements to their right shift values.

signedRightShift(n: number)

Parameters

n (number) Shift amount

▸ unsignedRightShift(n)

Set all elements to their unsigned right shift values.

unsignedRightShift(n: number)

Parameters

n (number) Shift amount

▸ broadcastOperate(o, fn)

Apply function for all elements with broadcasting.

broadcastOperate(o: (Matrix | Tensor | number), fn: function (number, number): number)

Parameters

o ((Matrix | Tensor | number)) Applied value

fn (function (number, number): number) Applied function

▸ operateAt(r, c, fn?)

Apply function to the position.

operateAt(r: (number | [number, number]), c: (number | function (number): number), fn: function (number): number?): number

Parameters

r ((number | [number, number])) Index of the row to apply function to

c ((number | function (number): number)) Index of the column to apply function to

fn (function (number): number?) Applied function

Returns

number: Old value

▸ add(o)

Add a value or matrix.

add(o: (Matrix | Tensor | number))

Parameters

o ((Matrix | Tensor | number)) Value to add

▸ addAt(r, c, v)

Add a value to the position.

addAt(r: number, c: number, v: number): number

Parameters

r (number) Index of the row to add the value to

c (number) Index of the column to add the value to

v (number) Value to add

Returns

number: Old value

▸ sub(o)

Subtract a value or matrix.

sub(o: (Matrix | Tensor | number))

Parameters

o ((Matrix | Tensor | number)) Value to subtract

▸ isub(o)

Subtract this matrix from a value or matrix.

isub(o: (Matrix | Tensor | number))

Parameters

o ((Matrix | Tensor | number)) Value to be subtracted

▸ subAt(r, c, v)

Subtract a value from the value at the position.

subAt(r: number, c: number, v: number): number

Parameters

r (number) Index of the row to subtract the value to

c (number) Index of the column to subtract the value to

v (number) Value to subtract

Returns

number: Old value

▸ isubAt(r, c, v)

Subtract the value at the position from a value.

isubAt(r: number, c: number, v: number): number

Parameters

r (number) Index of the row whose value is to be subtracted

c (number) Index of the column whose value is to be subtracted

v (number) Value to be subtracted

Returns

number: Old value

▸ mult(o)

Multiplies by a value element-wise.

mult(o: (Matrix | Tensor | number))

Parameters

o ((Matrix | Tensor | number)) Value to multiply

▸ multAt(r, c, v)

Multiplies a value to the position.

multAt(r: number, c: number, v: number): number

Parameters

r (number) Index of the row to multiply the value by

c (number) Index of the column to multiply the value by

v (number) Value to multiply

Returns

number: Old value

▸ div(o)

Divides by a value element-wise.

div(o: (Matrix | Tensor | number))

Parameters

o ((Matrix | Tensor | number)) Value to divide

▸ idiv(o)

Divides a value by this matrix element-wise.

idiv(o: (Matrix | Tensor | number))

Parameters

o ((Matrix | Tensor | number)) Value to be divided

▸ divAt(r, c, v)

Divides the value at the position by a value.

divAt(r: number, c: number, v: number): number

Parameters

r (number) Index of the row to divide the value by

c (number) Index of the column to divide the value by

v (number) Value to divide

Returns

number: Old value

▸ idivAt(r, c, v)

Divides a value by the value at the position.

idivAt(r: number, c: number, v: number): number

Parameters

r (number) Index of the row whose value is to be divided

c (number) Index of the column whose value is to be divided

v (number) Value to be divided

Returns

number: Old value

▸ mod(o)

Take a remainder divided by a value element-wise.

mod(o: (Matrix | Tensor | number))

Parameters

o ((Matrix | Tensor | number)) Value to divide

▸ imod(o)

Take a remainder divided a value by this matrix element-wise.

imod(o: (Matrix | Tensor | number))

Parameters

o ((Matrix | Tensor | number)) Value to be divided

▸ modAt(r, c, v)

Take a remainder divided the value at the position by a value.

modAt(r: number, c: number, v: number): number

Parameters

r (number) Index of the row to divide the value by

c (number) Index of the column to divide the value by

v (number) Value to divide

Returns

number: Old value

▸ imodAt(r, c, v)

Take a remainder divided a value by the value at the position.

imodAt(r: number, c: number, v: number): number

Parameters

r (number) Index of the row whose value is to be divided

c (number) Index of the column whose value is to be divided

v (number) Value to be divided

Returns

number: Old value

▸ and(o)

Take a logical AND with a value or matrix.

and(o: (Matrix | Tensor | number))

Parameters

o ((Matrix | Tensor | number)) Value to take a logical AND

▸ andAt(r, c, v)

Take logical AND with a value to the position.

andAt(r: number, c: number, v: number): number

Parameters

r (number) Index of the row to take a logical AND with

c (number) Index of the column to take a logical AND with

v (number) Value to take a logical AND

Returns

number: Old value

▸ or(o)

Take a logical OR with a value or matrix.

or(o: (Matrix | Tensor | number))

Parameters

o ((Matrix | Tensor | number)) Value to take a logical OR

▸ orAt(r, c, v)

Take logical OR with a value to the position.

orAt(r: number, c: number, v: number): number

Parameters

r (number) Index of the row to take a logical OR with

c (number) Index of the column to take a logical OR with

v (number) Value to take a logical OR

Returns

number: Old value

▸ bitand(o)

Take a bitwise AND with a value or matrix.

bitand(o: (Matrix | Tensor | number))

Parameters

o ((Matrix | Tensor | number)) Value to take a bitwise AND

▸ bitandAt(r, c, v)

Take bitwise AND with a value to the position.

bitandAt(r: number, c: number, v: number): number

Parameters

r (number) Index of the row to take a bitwise AND with

c (number) Index of the column to take a bitwise AND with

v (number) Value to take a bitwise AND

Returns

number: Old value

▸ bitor(o)

Take a bitwise OR with a value or matrix.

bitor(o: (Matrix | Tensor | number))

Parameters

o ((Matrix | Tensor | number)) Value to take a bitwise OR

▸ bitorAt(r, c, v)

Take bitwise OR with a value to the position.

bitorAt(r: number, c: number, v: number): number

Parameters

r (number) Index of the row to take a bitwise OR with

c (number) Index of the column to take a bitwise OR with

v (number) Value to take a bitwise OR

Returns

number: Old value

▸ bitxor(o)

Take a bitwise XOR with a value or matrix.

bitxor(o: (Matrix | Tensor | number))

Parameters

o ((Matrix | Tensor | number)) Value to take a bitwise XOR

▸ bitxorAt(r, c, v)

Take bitwise XOR with a value to the position.

bitxorAt(r: number, c: number, v: number): number

Parameters

r (number) Index of the row to take a bitwise XOR with

c (number) Index of the column to take a bitwise XOR with

v (number) Value to take a bitwise XOR

Returns

number: Old value

▸ dot(o)

Returns a matrix product value.

dot(o: Matrix): Matrix

Parameters

o (Matrix) Right matrix

Returns

Matrix: Producted matrix

▸ tDot(o)

Returns a matrix product of the transposed matrix of this and input.

tDot(o: Matrix): Matrix

Parameters

o (Matrix) Right matrix

Returns

Matrix: Producted matrix

▸ kron(mat)

Returns kronecker producted value.

kron(mat: Matrix): Matrix

Parameters

mat (Matrix) Right matrix

Returns

Matrix: Kronecker producted matrix

▸ convolute(kernel, normalize = true)

Convoluted by a kernel.

convolute(kernel: Array<Array<number>>, normalize: boolean?)

Parameters

kernel (Array<Array<number>>) Kernel matrix

normalize (boolean? = true) Normalize kernel or not

▸ reducedRowEchelonForm(tol = 0)

Calculate reduced row echelon form in-place.

reducedRowEchelonForm(tol: number?)

Parameters

tol (number? = 0) Tolerance to be recognized as 0

▸ inv()

Returns a inverse matrix.

inv(): Matrix

Returns

Matrix: Inversed matrix

▸ invLowerTriangular()

Returns a inverse matrix for lower triangular matrix.

invLowerTriangular(): Matrix

Returns

Matrix: Inversed matrix

▸ invUpperTriangular()

Returns a inverse matrix for upper triangular matrix.

invUpperTriangular(): Matrix

Returns

Matrix: Inversed matrix

▸ invRowReduction()

Returns a inverse matrix by row reduction.

invRowReduction(): Matrix

Returns

Matrix: Inversed matrix

▸ invLU()

Returns a inverse matrix by LU decompositioin.

invLU(): Matrix

Returns

Matrix: Inversed matrix

▸ pseudoInv()

Returns a pseudo inverse matrix.

pseudoInv(): Matrix

Returns

Matrix: pseudo inverse matrix

▸ pseudoInvNaive()

Returns a pseudo inverse matrix.

pseudoInvNaive(): Matrix

Returns

Matrix: pseudo inverse matrix

▸ pseudoInvQR()

Returns a Moore–Penrose inverse matrix by QR decomposition.

pseudoInvQR(): Matrix

Returns

Matrix: Moore–Penrose inverse matrix

▸ pseudoInvSVD()

Returns a Moore–Penrose inverse matrix by SVD decomposition.

pseudoInvSVD(): Matrix

Returns

Matrix: Moore–Penrose inverse matrix

▸ pseudoInvBenIsraelCohen()

Returns a Moore–Penrose inverse matrix by Ben-Israel and Cohen iterative method.

pseudoInvBenIsraelCohen(): Matrix

Returns

Matrix: Moore–Penrose inverse matrix

▸ sqrt()

Returns a square root of this matrix.

sqrt(): Matrix

Returns

Matrix: Squared matrix

▸ power(p)

Returns a power of this matrix.

power(p: number): Matrix

Parameters

p (number) Power exponent value

Returns

Matrix: Powered matrix

▸ exp()

Returns a exponential matrix

exp(): Matrix

Returns

Matrix: Exponential matrix

▸ log()

Returns a logarithm matrix

log(): Matrix

Returns

Matrix: Logarithm matrix

▸ cov(ddof = 0)

Returns a covariance matrix.

cov(ddof: number?): Matrix

Parameters

ddof (number? = 0) Delta Degrees of Freedom

Returns

Matrix: Covariance matrix

▸ gram()

Returns a gram matrix.

gram(): Matrix

Returns

Matrix: Gram matrix

▸ solve(b)

Returns a solved value A of a equation XA=B.

solve(b: Matrix): Matrix

Parameters

b (Matrix) Dependent variable values

Returns

Matrix: Solved matrix

▸ solveLowerTriangular(b)

Returns a solved value for lower triangular matrix.

solveLowerTriangular(b: Matrix): Matrix

Parameters

b (Matrix) Dependent variable values

Returns

Matrix: Solved matrix

▸ solveUpperTriangular(b)

Returns a solved value for upper triangular matrix.

solveUpperTriangular(b: Matrix): Matrix

Parameters

b (Matrix) Dependent variable values

Returns

Matrix: Solved matrix

▸ bidiag()

Returns a bidiagonal matrix.

bidiag(): Matrix

Returns

Matrix: Bidiagonal matrix

▸ bidiagHouseholder(return_uv = false)

Returns a bidiagonal matrix by Householder method.

bidiagHouseholder(return_uv: boolean?): (Matrix | [Matrix, Matrix, Matrix])

Parameters

return_uv (boolean? = false) Returns orthogonal matrixes

Returns

(Matrix | [Matrix, Matrix, Matrix]): Bidiagonal matrix, or Bidiagonal matrix and orthogonal matrixes

▸ tridiag()

Returns a tridiagonal matrix.

tridiag(): Matrix

Returns

Matrix: Tridiagonal matrix

▸ tridiagHouseholder(return_u = false)

Returns a tridiagonal matrix.

tridiagHouseholder(return_u: boolean?): (Matrix | [Matrix, Matrix])

Parameters

return_u (boolean? = false) Returns orthogonal matrix

Returns

(Matrix | [Matrix, Matrix]): Tridiagonal matrix, or Tridiagonal matrix and orthogonal matrix

▸ tridiagLanczos(k = 0)

Returns a tridiagonal matrix.

tridiagLanczos(k: number?): Matrix

Parameters

k (number? = 0) Number of iterations

Returns

Matrix: Tridiagonal matrix

▸ hessenberg()

Returns a hessenberg matrix.

hessenberg(): Matrix

Returns

Matrix: Hessenberg matrix

▸ hessenbergArnoldi(k = 0)

Returns a hessenberg matrix.

hessenbergArnoldi(k: number?): Matrix

Parameters

k (number? = 0) Number of iterations

Returns

Matrix: Hessenberg matrix

▸ balancing()

Returns a doubly stochastic matrix.

balancing(): [Array<number>, Matrix, Array<number>]

Returns

[Array<number>, Matrix, Array<number>]: Doubly stochastic matrix

▸ balancingSinkhornKnopp()

Returns a doubly stochastic matrix by Sinkhorn-Knopp algorithm.

balancingSinkhornKnopp(): [Array<number>, Matrix, Array<number>]

Returns

[Array<number>, Matrix, Array<number>]: Doubly stochastic matrix

▸ lu()

Returns a LU decomposition.

lu(): [Matrix, Matrix]

Returns

[Matrix, Matrix]: Lower triangular matrix and upper triangular matrix

▸ qr()

Returns a QR decomposition.

qr(): [Matrix, Matrix]

Returns

[Matrix, Matrix]: Orthogonal matrix and upper triangular matrix

▸ qrGramSchmidt()

Returns a QR decomposition by Gram-Schmidt method.

qrGramSchmidt(): [Matrix, Matrix]

Returns

[Matrix, Matrix]: Orthogonal matrix and upper triangular matrix

▸ qrHouseholder()

Returns a QR decomposition by Householder method.

qrHouseholder(): [Matrix, Matrix]

Returns

[Matrix, Matrix]: Orthogonal matrix and upper triangular matrix

▸ singularValues()

Returns singular values.

singularValues(): Array<number>

Returns

Array<number>: Singular values

▸ svd()

Returns a singular value decomposition.

svd(): [Matrix, Array<number>, Matrix]

Returns

[Matrix, Array<number>, Matrix]: Unitary matrix and singular values

▸ svdEigen()

Returns a singular value decomposition by eigen decomposition.

svdEigen(): [Matrix, Array<number>, Matrix]

Returns

[Matrix, Array<number>, Matrix]: Unitary matrix and singular values

▸ svdGolubKahan()

Returns a singular value decomposition by Golub-Kahan method.

svdGolubKahan(): [Matrix, Array<number>, Matrix]

Returns

[Matrix, Array<number>, Matrix]: Unitary matrix and singular values

▸ cholesky()

Returns a cholesky decomposition.

cholesky(): Matrix

Returns

Matrix: Cholesky decomposition matrix

▸ choleskyGaussian()

Returns a cholesky decomposition by Gaussian algorithm.

choleskyGaussian(): Matrix

Returns

Matrix: Cholesky decomposition matrix

▸ choleskyBanachiewicz()

Returns a cholesky decomposition by Banachiewicz algorithm.

choleskyBanachiewicz(): Matrix

Returns

Matrix: Cholesky decomposition matrix

▸ choleskyCrout()

Returns a cholesky decomposition by Crout algorithm.

choleskyCrout(): Matrix

Returns

Matrix: Cholesky decomposition matrix

▸ modifiedCholesky()

Returns a modified cholosky decomposition.

modifiedCholesky(): [Matrix, Array<number>]

Returns

[Matrix, Array<number>]: Cholesky decomposition matrix and diagonal matrix

▸ schur()

Returns schur decomposition.

schur(): [Matrix, Matrix]

Returns

[Matrix, Matrix]: Schur decomposition matrix

▸ schurQR(shift = 'single')

Returns schur decomposition by QR decomposition.

schurQR(shift: ("no" | "single")?): [Matrix, Matrix]

Parameters

shift (("no" | "single")? = 'single') Shifting type

Returns

[Matrix, Matrix]: Schur decomposition matrix

▸ rankFactorization()

Returns rank factorization.

rankFactorization(): [Matrix, Matrix]

Returns

[Matrix, Matrix]: Rank factorization matrix

▸ eigen()

Returns eigenvalues and eigenvectors.

eigen(): [Array<number>, Matrix]

Returns

[Array<number>, Matrix]: Eigenvalues and eigenvectors. Eigenvectors correspond to each column of the matrix.

▸ eigenValues()

Returns eigenvalues.

eigenValues(): Array<number>

Returns

Array<number>: Eigenvalues

▸ eigenVectors()

Returns eigenvectors.

eigenVectors(): Matrix

Returns

Matrix: Eigenvectors. Eigenvectors correspond to each column of the matrix.

▸ eigenValuesBiSection()

Returns eigenvalues by Bi-section.

eigenValuesBiSection(): Array<number>

Returns

Array<number>: Eigenvalues

▸ eigenValuesLR(maxIteration = 1.0e5)

Returns eigenvalues by LU decomposition.

eigenValuesLR(maxIteration: number?): Array<number>

Parameters

maxIteration (number? = 1.0e5) Maximum iteration

Returns

Array<number>: Eigenvalues

▸ eigenValuesQR(maxIteration = 1.0e6)

Returns eigenvalues by QR decomposition.

eigenValuesQR(maxIteration: number?): Array<number>

Parameters

maxIteration (number? = 1.0e6) Maximum iteration

Returns

Array<number>: Eigenvalues

▸ eigenJacobi(maxIteration = 1.0e6)

Returns eigenvalues and eigenvectors.

eigenJacobi(maxIteration: number?): [Array<number>, Matrix]

Parameters

maxIteration (number? = 1.0e6) Maximum iteration

Returns

[Array<number>, Matrix]: Eigenvalues and eigenvectors. Eigenvectors correspond to each column of the matrix.

▸ eigenPowerIteration(maxIteration = 1.0e4)

Returns the maximum eigenvalue and its eigenvector.

eigenPowerIteration(maxIteration: number?): [number, Matrix]

Parameters

maxIteration (number? = 1.0e4) Maximum iteration

Returns

[number, Matrix]: Maximum eigenvalue and its eigenvector

▸ eigenInverseIteration(ev = 0.0, maxIteration = 1.0e4)

Returns the nearest eigenvalue and its eigenvector to the specified value.

eigenInverseIteration(ev: number?, maxIteration: number?): [number, Matrix]

Parameters

ev (number? = 0.0) Target value

maxIteration (number? = 1.0e4) Maximum iteration

Returns

[number, Matrix]: Eigenvalue and eigenvector

Tensor

Tensor class

new Tensor(size: Array<number>, value: (number | Array<number>)?)

Parameters

size (Array<number>) Sizes for each dimension

value ((number | Array<number>)?) Initial values

Static Members

▸ zeros(size)

Returns a tensor filled with 0.

zeros(size: Array<number>): Tensor

Parameters

size (Array<number>) Sizes for each dimension

Returns

Tensor: Tensor filled with 0

▸ ones(size)

Returns a tensor filled with 1.

ones(size: Array<number>): Tensor

Parameters

size (Array<number>) Sizes for each dimension

Returns

Tensor: Tensor filled with 1

▸ random(size, min = 0, max = 1)

Returns a tensor initialized uniform random values.

random(size: Array<number>, min: number?, max: number?): Tensor

Parameters

size (Array<number>) Sizes for each dimension

min (number? = 0) Minimum value of the Tensor

max (number? = 1) Maximum value of the Tensor

Returns

Tensor: Tensor initialized uniform random values

▸ randn(size, myu = 0, sigma = 1)

Returns a tensor initialized normal random values.

randn(size: Array<number>, myu: number?, sigma: number?): Tensor

Parameters

size (Array<number>) Sizes for each dimension

myu (number? = 0) Mean value of the Tensor

sigma (number? = 1) Variance value of the Tensor

Returns

Tensor: Tensor initialized normal random values

▸ fromArray(arr)

Returns a tensor from some value.

fromArray(arr: (Tensor | Matrix | Array<number> | number)): Tensor

Parameters

arr ((Tensor | Matrix | Array<number> | number)) Original values

Returns

Tensor: Tensor from some value

Instance Members

▸ dimension

Dimension of the tensor.

dimension

Type: number

▸ sizes

Sizes of the tensor.

sizes

Type: Array<number>

▸ length

Number of all elements in the tensor.

length

Type: number

▸ value

Elements in the tensor.

value

Type: Array<number>

▸ iterator()

Iterate over the elements.

iterator()

▸ toArray()

Returns a nested array represented this tensor.

toArray(): Array<number>

Returns

Array<number>: Nested array

▸ toString()

Returns a string represented this tensor.

toString(): string

Returns

string: String represented this tensor

▸ toMatrix()

Returns a Matrix if the dimension of this tensor is 2.

toMatrix(): Matrix

Returns

Matrix: Matrix

Throws

MatrixException: If the dimension of this tensor is not 2.

▸ toScaler()

Returns the only element.

toScaler(): number

Returns

number: The only element

▸ copy()

Returns a copy of this tensor.

copy(): Tensor

Returns

Tensor: Copied tensor

▸ equals(other)

Returns this tensor is equals to the others.

equals(other: any): boolean

Parameters

other (any) Check tensor

Returns

boolean: true if equal

▸ at(i)

Returns value at the index position.

at(i: ...number): number

Parameters

i (...number) Index values

Returns

number: The value

▸ index(i)

Returns tensor at the index position.

index(i: ...number): Tensor

Parameters

i (...number) Index values

Returns

Tensor: Sub tensor

▸ set(i, value)

Set the value at the specific position.

set(i: Array<number>, value: number)

Parameters

i (Array<number>) Index values

value (number) Set value

▸ select(idx, axis = 0)

Returns the sub-tensor corresponding to position i in the first dimension of this.

select(idx: (number | Array<number>), axis: number?): Tensor

Parameters

idx ((number | Array<number>)) Select index value(s)

axis (number? = 0) Axis

Returns

Tensor: Selected tensor

▸ slice(from, to, axis = 0)

Returns a tensor sliced by first dimension.

slice(from: number, to: number, axis: number?): Tensor

Parameters

from (number) Start index

to (number) End index

axis (number? = 0) Axis

Returns

Tensor: Sliced tensor

▸ fill(value)

Fill in all the elements with the value.

fill(value: number)

Parameters

value (number) Filled value

▸ map(cb)

Iterate over all the elements and replace the value.

map(cb: function (number, Array<number>, Tensor): number)

Parameters

cb (function (number, Array<number>, Tensor): number) Mapping function

▸ forEach(cb)

Iterate over all the elements.

forEach(cb: function (number, Array<number>, Tensor): void)

Parameters

cb (function (number, Array<number>, Tensor): void) Callback function

▸ transpose(axises)

Returns a tensor transposed along the axis.

transpose(axises: ...number): Tensor

Parameters

axises (...number) Selected axises

Returns

Tensor: Transposed tensor

▸ flip(axis = 0)

Flip values along the axis.

flip(axis: number?)

Parameters

axis (number? = 0) Axis to be flipped

▸ shuffle(axis = 0)

Shuffle along the axis.

shuffle(axis: number?)

Parameters

axis (number? = 0) Axis

▸ resize(sizes, init = 0)

Resize this tensor.

resize(sizes: Array<number>, init: number?)

Parameters

sizes (Array<number>) New sizes

init (number? = 0) Value of the extended region

▸ reshape(sizes)

Reshape this as the sizes.

reshape(sizes: ...number)

Parameters

sizes (...number) New sizes for each dimension

▸ repeat(n, axis = 0)

Repeat the elements n times along the axis this.

repeat(n: (number | Array<number>), axis: number?)

Parameters

n ((number | Array<number>)) Repeated count(s)

axis (number? = 0) Axis to be repeated

▸ concat(t, axis = 0)

Concatenate this and t.

concat(t: Tensor, axis: number?)

Parameters

t (Tensor) Concatenate tensor

axis (number? = 0) Axis to be concatenated

▸ reduce(cb, init?, axis = -1, keepdims = false)

Returns a tensor reduced along the axis with the callback function.

reduce(cb: function (number, number, Array<number>, Tensor): number, init: any?, axis: (number | Array<number>)?, keepdims: boolean?): (Tensor | number)

Parameters

cb (function (number, number, Array<number>, Tensor): number) Reducing function

init (any?) Initial value

axis ((number | Array<number>)? = -1) Axis to be reduced. If negative, reduce along all elements.

keepdims (boolean? = false) Keep dimensions or not.

Returns

(Tensor | number): Reduced tensor or value

▸ broadcastOperate(o, fn)

Apply function for all elements with broadcasting.

broadcastOperate(o: (Tensor | Matrix | number), fn: function (number, number): number)

Parameters

o ((Tensor | Matrix | number)) Applied value

fn (function (number, number): number) Applied function

▸ operateAt(i, fn?)

Apply function to the position.

operateAt(i: Array<number>, fn: function (number): number?): number

Parameters

i (Array<number>) Index values

fn (function (number): number?) Applied function

Returns

number: Old value

▸ dot(o)

Returns a tensor product value.

dot(o: Matrix): Tensor

Parameters

o (Matrix) Right matrix

Returns

Tensor: Producted tensor

GraphException

Exception for graph class

new GraphException(message: string, value: any)

Extends Error

Parameters

message (string) Error message

value (any) Some value

Edge

Edge of graph

new Edge(from: number, to: number, value: unknown?, direct: boolean?)

Parameters

from (number) Index of the starting node of the edge

to (number) Index of the end node of the edge

value (unknown? = null) Value of the edge

direct (boolean? = false) true if the edge is direct

Graph

Graph class

new Graph(nodes: (number | Array<unknown>)?, edges: any)

Parameters

nodes ((number | Array<unknown>)? = 0) Number of nodes or values of nodes

edges (any = [])

Static Members

▸ fromAdjacency(mat)

Returns graph from adjacency matrix.

fromAdjacency(mat: any): Graph

Parameters

mat (any)

Returns

Graph: Graph from adjacency matrix

▸ complete(n)

Returns complete graph.

complete(n: number): Graph

Parameters

n (number) Size of the graph

Returns

Graph: Complete graph

▸ completeBipartite(n, m)

Returns complete bipartite graph.

completeBipartite(n: number, m: number): Graph

Parameters

n (number) Size of the first group

m (number) Size of the second group

Returns

Graph: Complete bipartite graph

▸ cycle(n, direct = false)

Returns cycle graph.

cycle(n: number, direct: boolean?): Graph

Parameters

n (number) Size of the graph

direct (boolean? = false) Direct graph or not

Returns

Graph: Cycle graph

▸ wheel(n)

Returns wheel graph.

wheel(n: number): Graph

Parameters

n (number) Size of the graph

Returns

Graph: Wheel graph

▸ windmill(k, n)

Returns windmill graph.

windmill(k: number, n: number): Graph

Parameters

k (number) Size of the sub complete graph

n (number) Number of the sub complete graph

Returns

Graph: Windmill graph

▸ fromName(name)

Returns named graph

Parameters

name

(("balaban 10 cage" | "bidiakis cube" | "biggs smith" | "brinkmann" | "bull" | "butterfly" | "chvatal" | "clebsch" | "coxeter" | "desargues" | "diamond" | "durer" | "errera" | "folkman" | "foster" | "franklin" | "frucht" | "goldner-harary" | "golomb" | "gray" | "grotzsch" | "harries" | "heawood" | "herschel" | "hoffman" | "holt" | "kittell" | "markstrom" | "mcgee" | "meredith" | "mobius kantor" | "moser spindle" | "nauru" | "pappus" | "petersen" | "poussin" | "robertson" | "shrikhande" | "sousselier" | "sylvester" | "tutte" | "tutte coxeter" | "wagner" | "wells"))

Name of the graph

Returns

Graph: Named graph

Instance Members

▸ order

Number of nodes

order

Type: number

▸ size

Number of edges

size

Type: number

▸ nodes

Nodes

nodes

Type: Array<unknown>

▸ edges

Edges

edges

Type: Array<Edge>

▸ toDot()

Returns a string of DOT format.

toDot(): string

Returns

string: String of DOT format

▸ toString()

Returns a string represented this graph.

toString(): string

Returns

string: String represented this graph

▸ copy()

Returns a copy of this graph.

copy(): Graph

Returns

Graph: Copied grpah

▸ degree(k, undirect = true, direct = true)

Return degree of the node.

degree(k: number, undirect: (boolean | "in" | "out")?, direct: (boolean | "in" | "out")?): number

Parameters

k (number) Index of target node

undirect ((boolean | "in" | "out")? = true) Count undirected edges. If in or out is specified, only direct edges are counted and direct parameter is ignored.

direct ((boolean | "in" | "out")? = true) Count directed edges

Returns

number: Degree of the node

▸ adjacencies(k, undirect = true, direct = true)

Return indexes of adjacency nodes.

adjacencies(k: number, undirect: (boolean | "in" | "out")?, direct: (boolean | "in" | "out")?): Array<number>

Parameters

k (number) Index of target node

undirect ((boolean | "in" | "out")? = true) Check undirected edges. If in or out is specified, only direct edges are checked and direct parameter is ignored.

direct ((boolean | "in" | "out")? = true) Check directed edges

Returns

Array<number>: Indexes of adjacency nodes

▸ components()

Returns indexes of each components.

components()

▸ biconnectedComponents()

Returns indexes of each biconnected components.

biconnectedComponents()

▸ diameter()

Returns diameter of this graph.

diameter(): number

Returns

number: Diameter

▸ eccentricity(k)

Returns eccentricity at k of this graph.

eccentricity(k: number): number

Parameters

k (number) Index of target node

Returns

number: Eccentricity

▸ radius()

Returns radius of this graph.

radius(): number

Returns

number: Radius

▸ center()

Returns indexes of center of this graph.

center(): Array<number>

Returns

Array<number>: Indexes of center

▸ girth()

Returns girth of this graph.

girth(): number

Returns

number: Girth

▸ clique(k?)

Returns index of cliques.

clique(k: number?)

Parameters

k (number?) Size of clique

▸ chromaticNumber()

Returns chromatic number of this graph.

chromaticNumber(): number

Returns

number: Chromatic number

▸ chromaticNumberWelchPowell()

Returns chromatic number of this graph with Welch-Powell algorithm.

chromaticNumberWelchPowell(): number

Returns

number: Chromatic number

▸ chromaticIndex()

Returns chromatic index of this graph.

chromaticIndex(): number

Returns

number: Chromatic index

▸ articulations()

Returns indexes of articulation (cut) nodes.

articulations(): Array<number>

Returns

Array<number>: Indexes of articulation nodes

▸ articulationsEachNodes()

Returns indexes of articulation (cut) nodes with checking each node.

articulationsEachNodes(): Array<number>

Returns

Array<number>: Indexes of articulation nodes

▸ articulationsLowLink()

Returns indexes of articulation (cut) nodes with checking lowlinks.

articulationsLowLink(): Array<number>

Returns

Array<number>: Indexes of articulation nodes

▸ bridges()

Returns edges of bridge.

bridges(): Array<Edge>

Returns

Array<Edge>: Bridge edges

▸ bridgesLowLink()

Returns edges of bridge with checking lowlinks.

bridgesLowLink(): Array<Edge>

Returns

Array<Edge>: Bridge edges

▸ addNode(value?)

Add the node.

addNode(value: unknown?)

Parameters

value (unknown?) Value of the node

▸ getNode(k?)

Returns the node value.

getNode(k: (number | Array<number>)?): (unknown | Array<unknown>)

Parameters

k ((number | Array<number>)?) Index of the node

Returns

(unknown | Array<unknown>): Node value

▸ removeNode(k)

Remove the node.

removeNode(k: number)

Parameters

k (number) Index of the node

▸ clearNodes()

Remove all nodes.

clearNodes()

▸ addEdge(from, to, value = null, direct = false)

Add the edge.

addEdge(from: number, to: number, value: unknown?, direct: boolean?)

Parameters

from (number) Index of the starting node of the edge

to (number) Index of the end node of the edge

value (unknown? = null) Value of the edge

direct (boolean? = false) true if the edge is direct

▸ getEdges(from, to, undirect = true, direct = true)

Returns the edges.

getEdges(from: number, to: number, undirect: (boolean | "forward" | "backward")?, direct: (boolean | "forward" | "backward")?): Array<Edge>

Parameters

from (number) Index of the starting node of the edge

to (number) Index of the end node of the edge

undirect ((boolean | "forward" | "backward")? = true) Get undirected edges or not. If forward or backward is specified, only direct edges are get and direct parameter is ignored.

direct ((boolean | "forward" | "backward")? = true) Get directed edges or not

Returns

Array<Edge>: Edges between from and to

▸ removeEdges(from, to, direct = null)

Remove the edges.

removeEdges(from: number, to: number, direct: (boolean | null)?)

Parameters

from (number) Index of the starting node of the edge

to (number) Index of the end node of the edge

direct ((boolean | null)? = null) null to remove direct and undirect edges, true to remove only direct edges, false to remove only undirect edges.

▸ clearEdges()

Remove all edges.

clearEdges()

▸ adjacencyMatrix()

Returns adjacency matrix

adjacencyMatrix()

▸ adjacencyList(direct = 'both')

Returns adjacency list

adjacencyList(direct: ("both" | "in" | "out")?)

Parameters

direct (("both" | "in" | "out")? = 'both') Indegree or outdegree

▸ degreeMatrix(direct = 'both')

Returns degree matrix.

degreeMatrix(direct: ("both" | "in" | "out")?)

Parameters

direct (("both" | "in" | "out")? = 'both') Indegree or outdegree

▸ laplacianMatrix()

Returns laplacian matrix.

laplacianMatrix()

▸ isNull()

Returns if this is null graph or not.

isNull(): boolean

Returns

boolean: true if this is null graph

▸ isEdgeless()

Returns if this is edgeless graph or not.

isEdgeless(): boolean

Returns

boolean: true if this is edgeless graph

▸ isUndirected()

Returns if this is undirected graph or not.

isUndirected(): boolean

Returns

boolean: true if this is undirected graph

▸ isDirected()

Returns if this is directed graph or not.

isDirected(): boolean

Returns

boolean: true if this is directed graph

▸ isMixed()

Returns if this is mixed graph or not.

isMixed(): boolean

Returns

boolean: true if this is mixed graph

▸ isOriented()

Returns if this is oriented graph or not.

isOriented(): boolean

Returns

boolean: true if this is oriented graph

▸ isWeighted()

Returns if this is weighted graph or not.

isWeighted(): boolean

Returns

boolean: true if this is weighted graph

▸ isSimple()

Returns if this is simple graph or not.

isSimple(): boolean

Returns

boolean: true if this is simple graph

▸ isConnected()

Returns if this is connected graph or not.

isConnected(): boolean

Returns

boolean: true if this is connected graph

▸ isBiconnected()

Returns if this is biconnected graph or not.

isBiconnected(): boolean

Returns

boolean: true if this is biconnected graph

▸ isTree()

Returns if this is tree or not.

isTree(): boolean

Returns

boolean: true if this is tree

▸ isForest()

Returns if this is forest or not.

isForest(): boolean

Returns

boolean: true if this is forest

▸ isBipartite()

Returns if this is bipartite graph or not.

isBipartite(): boolean

Returns

boolean: true if this is bipartite graph

▸ isComplete()

Returns if this is complete graph or not.

isComplete(): boolean

Returns

boolean: true if this is complete graph

▸ isRegular(n = null)

Returns if this is regular graph or not.

isRegular(n: number?): boolean

Parameters

n (number? = null) Degree of vertices

Returns

boolean: true if this is regular graph

▸ isPlainer()

Returns if this is plainer graph or not.

isPlainer(): boolean

Returns

boolean: true if this is plainer graph

▸ isPlainerAddEdge()

Returns if this is plainer graph or not with add-path algorithm.

On the Cutting Edge: Simplified O(n) Planarity by Edge Addition https://xuzijian629.hatenablog.com/entry/2019/12/14/163726

isPlainerAddEdge(): boolean

Returns

boolean: true if this is plainer graph

▸ isPlainerAddVertex()

Returns if this is plainer graph or not with add-vertex algorithm.

Hopcroft, J. and Tarjan, R. "Efficient Planarity Testing", J. ACM, Vol. 21, No. 4, pp. 549-568 (1974) 西関隆夫. "32. グラフの平面性判定法", 情報処理, Vol. 24, No. 4, pp. 521-528 (1983) K. S. Booth, "Testing for the Consecutive Ones Property, Interval Graphs, and Graph Planarity Using PQ-Tree Algorithms", Journal of computer and system sciences, 13, pp. 335-379 (1976)

isPlainerAddVertex(): boolean

Returns

boolean: true if this is plainer graph

▸ isSymmetric()

Returns if this is symmetric graph or not.

isSymmetric(): boolean

Returns

boolean: true if this is symmetric graph

▸ isDAG()

Returns if this is directed acyclic graph or not.

isDAG(): boolean

Returns

boolean: true if this is directed acyclic graph

▸ isSeparable()

Returns if this is separable graph or not.

isSeparable(): boolean

Returns

boolean: true if this is separable graph

▸ isEulerian()

Returns if this is Eulerian graph or not.

isEulerian(): boolean

Returns

boolean: true if this is Eulerian graph

▸ isSemiEulerian()

Returns if this is semi-Eulerian graph or not.

isSemiEulerian(): boolean

Returns

boolean: true if this is semi-Eulerian graph

▸ isHamiltonian()

Returns if this is Hamiltonian graph or not.

isHamiltonian(): boolean

Returns

boolean: true if this is Hamiltonian graph

▸ isSemiHamiltonian()

Returns if this is semi-Hamiltonian graph or not.

isSemiHamiltonian(): boolean

Returns

boolean: true if this is semi-Hamiltonian graph

▸ hasCycle()

Returns if this has cycle or not.

hasCycle(): boolean

Returns

boolean: true if this has cycle

▸ hasCycleDFS()

Returns if this has cycle or not with depth-first search.

hasCycleDFS(): boolean

Returns

boolean: true if this has cycle

▸ hasCycleEachNodes()

Returns if this has cycle or not with checking each node.

hasCycleEachNodes(): boolean

Returns

boolean: true if this has cycle

▸ toUndirected()

Returns graph of directed edges converted to undirected.

toUndirected(): Graph

Returns

Graph: Undirected graph

▸ toSimple()

Returns a graph with multiple edges and loops removed.

toSimple(): Graph

Returns

Graph: Simple graph

▸ isomorphism(g)

Returns (sub) graph isomorphism maps from 'g' to this (sub) graph.

isomorphism(g: Graph)

Parameters

g (Graph) Other graph

▸ isomorphismUllmann(g)

Returns (sub) graph isomorphism maps from 'g' to this (sub) graph with Ullmann algorithm.

isomorphismUllmann(g: Graph)

Parameters

g (Graph) Other graph

▸ isomorphismVF2(g)

Returns (sub) graph isomorphism maps from 'g' to this (sub) graph with VF2 algorithm.

isomorphismVF2(g: Graph)

Parameters

g (Graph) Other graph

▸ inducedSub(k)

Returns induced sub graph.

inducedSub(k: Array<number>): Graph

Parameters

k (Array<number>) Selected indexes

Returns

Graph: Induced sub graph

▸ complement()

Returns complement graph.

complement(): Graph

Returns

Graph: Complement graph

▸ line()

Returns line graph.

line(): Graph

Returns

Graph: Line graph

▸ contraction(a, b)

Contract this graph.

contraction(a: number, b: number)

Parameters

a (number) Index of node

b (number) Index of node

▸ subdivision(a, b)

Subdivision this graph.

subdivision(a: number, b: number)

Parameters

a (number) Index of node

b (number) Index of node

▸ cleaving(a)

Cleave the node.

cleaving(a: number)

Parameters

a (number) Index of node

▸ disjointUnion(g)

Take the disjoint union of this graph and other graph.

disjointUnion(g: Graph)

Parameters

g (Graph) Other graph

▸ substitution(k, g)

Substitute other graph at the node.

substitution(k: number, g: Graph)

Parameters

k (number) Index of the node

g (Graph) Other graph

▸ cartesianProduct(g)

Take the cartesian product of this graph and other graph.

cartesianProduct(g: Graph): Graph

Parameters

g (Graph) Other graph

Returns

Graph: Cartesian producted graph

▸ tensorProduct(g)

Take the tensor product of this graph and other graph.

tensorProduct(g: Graph): Graph

Parameters

g (Graph) Other graph

Returns

Graph: Tensor producted graph

▸ strongProduct(g)

Take the strong product of this graph and other graph.

strongProduct(g: Graph): Graph

Parameters

g (Graph) Other graph

Returns

Graph: Strong producted graph

▸ lexicographicProduct(g)

Take the lexicographic product of this graph and other graph.

lexicographicProduct(g: Graph): Graph

Parameters

g (Graph) Other graph

Returns

Graph: Lexicographic producted graph

▸ shortestPath(from?)

Returns shortest path.

shortestPath(from: number?)

Parameters

from (number?) Index of start nodes

▸ shortestPathBreadthFirstSearch(from)

Returns shortest path with breadth first search algorithm.

shortestPathBreadthFirstSearch(from: number): Array<{length: number, prev: number, path: Array<number>}>

Parameters

from (number) Index of start node

Returns

Array<{length: number, prev: number, path: Array<number>}>: Shortest length and path for all nodes

▸ shortestPathDijkstra(from)

Returns shortest path with Dijkstra's algorithm.

shortestPathDijkstra(from: number): Array<{length: number, prev: number, path: Array<number>}>

Parameters

from (number) Index of start node

Returns

Array<{length: number, prev: number, path: Array<number>}>: Shortest length and path for all nodes

▸ shortestPathBellmanFord(from)

Returns shortest path with Bellman–Ford algorithm.

shortestPathBellmanFord(from: number): Array<{length: number, prev: number, path: Array<number>}>

Parameters

from (number) Index of start node

Returns

Array<{length: number, prev: number, path: Array<number>}>: Shortest length and path for all nodes

▸ shortestPathFloydWarshall()

Returns shortest path with Floyd–Warshall algorithm.

shortestPathFloydWarshall()

▸ minimumSpanningTree()

Returns minimum spanning tree.

minimumSpanningTree(): Graph

Returns

Graph: Minimum spanning tree

▸ minimumSpanningTreePrim()

Returns minimum spanning tree with Prim's algorithm.

minimumSpanningTreePrim(): Graph

Returns

Graph: Minimum spanning tree

▸ minimumSpanningTreeKruskal()

Returns minimum spanning tree with Kruskal's algorithm.

minimumSpanningTreeKruskal(): Graph

Returns

Graph: Minimum spanning tree

▸ minimumSpanningTreeBoruvka()

Returns minimum spanning tree with Borůvka's algorithm.

minimumSpanningTreeBoruvka(): Graph

Returns

Graph: Minimum spanning tree

▸ hamiltonianPath(from?)

Returns Hamiltonian path

hamiltonianPath(from: number?)

Parameters

from (number?) Index of start node

▸ hamiltonianPathDynamicProgramming(from?)

Returns Hamiltonian path with dynamic programming

hamiltonianPathDynamicProgramming(from: number?)

Parameters

from (number?) Index of start node

▸ hamiltonianCycle()

Returns Hamiltonian cycle

hamiltonianCycle()

▸ cut(s, t)

Returns cut size.

cut(s: Array<number>, t: Array<number>): number

Parameters

s (Array<number>) Subset

t (Array<number>) Subset

Returns

number: Cut size

▸ mincut(minv = 1)

Returns minimum cut.

mincut(minv: number?)

Parameters

minv (number? = 1) Minimum number for subset

▸ mincutBruteForce(minv = 1)

Returns minimum cut.

mincutBruteForce(minv: number?)

Parameters

minv (number? = 1) Minimum number for subset

▸ mincutStoerWagner(minv = 1, startnode = 0)

Returns minimum cut.

mincutStoerWagner(minv: number?, startnode: number?)

Parameters

minv (number? = 1) Minimum number for subset

startnode (number? = 0) Start node index

▸ mincutKargers(minv = 1, trials = null)

Returns minimum cut.

mincutKargers(minv: number?, trials: number?)

Parameters

minv (number? = 1) Minimum number for subset

trials (number? = null) Trial count

▸ mincutKargersStein(minv = 1, trials = null)

Returns minimum cut.

mincutKargersStein(minv: number?, trials: number?)

Parameters

minv (number? = 1) Minimum number for subset

trials (number? = null) Trial count

▸ bisectionSpectral()

Returns bisection cut.

bisectionSpectral()

Complex

Complex number

new Complex(real: number?, imag: number?)

Parameters

real (number? = 0) Real number

imag (number? = 0) Imaginary number

Instance Members

▸ real

Real value.

real

Type: number

▸ imaginary

Imaginary value.

imaginary

Type: number

▸ abs()

Returns absolute value.

abs(): number

Returns

number: Absolute number

▸ conjugate()

Returns conjugate value.

conjugate(): Complex

Returns

Complex: Conjugate number

▸ add(other)

Returns added value.

add(other: (number | Complex)): Complex

Parameters

other ((number | Complex)) Number to add

Returns

Complex: Added complex number

▸ sub(other)

Returns subtracted value.

sub(other: (number | Complex)): Complex

Parameters

other ((number | Complex)) Number to subtract

Returns

Complex: Subtracted complex number

▸ mult(other)

Returns multiplicated value.

mult(other: (number | Complex)): Complex

Parameters

other ((number | Complex)) Number to multiplicate

Returns

Complex: Multiplicated complex number

▸ div(other)

Returns divided value.

div(other: (number | Complex)): Complex

Parameters

other ((number | Complex)) Number to divide

Returns

Complex: Divided complex number

▸ sqrt()

Returns sqare root values.

sqrt(): [Complex, Complex]

Returns

[Complex, Complex]: Sqare root complex numbers

▸ cbrt()

Returns cubic root values.

cbrt(): [Complex, Complex, Complex]

Returns

[Complex, Complex, Complex]: Cubic root complex numbers

▸ exp()

Returns value of complex exponential function.

exp(): Complex

Returns

Complex: Exponential value

▸ log()

Returns value of complex log function.

log(): Complex

Returns

Complex: Principal log value

A2CAgent

A2C agent

new A2CAgent(env: RLEnvironmentBase, resolution: number, procs: number, layers: Array<LayerObject>, optimizer: string)

Parameters

env (RLEnvironmentBase) Environment

resolution (number) Resolution of actions

procs (number) Number of processes

layers (Array<LayerObject>) Network layers

optimizer (string) Optimizer of the network

Instance Members

▸ get_score()

Returns a score.

get_score(): Array<Array<Array<number>>>

Returns

Array<Array<Array<number>>>: Score values

▸ get_action(state)

Returns a action.

get_action(state: Array<any>): Array<any>

Parameters

state (Array<any>) Current states

Returns

Array<any>: Action

▸ update(done, learning_rate, batch)

Update model.

update(done: boolean, learning_rate: number, batch: number): number

Parameters

done (boolean) Done epoch or not

learning_rate (number) Learning rate

batch (number) Batch size

Returns

number: Loss value

ABOD

Angle-based Outlier Detection

new ABOD(k: number?)

Parameters

k (number? = Infinity) Number of neighborhoods

Instance Members

▸ predict(datas)

Returns anomaly degrees.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Predicted values

LBABOD

Lower-bound for the Angle-based Outlier Detection

new LBABOD(k: number?, l: number?)

Parameters

k (number? = 10) Number of neighborhoods

l (number? = 5) Number of outliers

Instance Members

▸ predict(datas)

lib/model/adaptive_thresholding.js

Returns a list of the data predicted as outliers or not.

predict(datas: Array<Array<number>>): Array<boolean>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<boolean>: Predicted values

ADALINE

lib/model/adaline.js

Adaptive Linear Neuron model

new ADALINE(rate: number)

Parameters

rate (number) Learning rate

Instance Members

▸ fit(x, y)

lib/model/adaline.js

Fit this model once.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(data)

lib/model/adaline.js

Returns predicted datas.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

ADAMENN

lib/model/adamenn.js

Adaptive Metric Nearest Neighbor

new ADAMENN(k0: number?, k1: number?, k2: number?, l: number?, k: number?, c: number?)

Parameters

k0 (number? = null) The number of neighbors of the test point

k1 (number? = 3) The number of neighbors in N1 for estimation

k2 (number? = null) The size of the neighborhood N2 for each of the k0 neighbors for estimation

l (number? = null) The number of points within the delta intervals

k (number? = 3) The number of neighbors in the final nearest neighbor rule

c (number? = 0.5) The positive factor for the exponential weighting scheme

Instance Members

▸ fit(x, y)

lib/model/adamenn.js

Fit model.

fit(x: Array<Array<number>>, y: Array<any>)

Parameters

x (Array<Array<number>>) Training data

y (Array<any>) Target values

▸ predict(datas)

lib/model/adamenn.js

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<any>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

AdaptiveThresholding

Adaptive thresholding

new AdaptiveThresholding(method: ("mean" | "gaussian" | "median" | "midgray"), k: number, c: number)

Parameters

method (("mean" | "gaussian" | "median" | "midgray") = 'mean') Method name

k (number = 3) Size of local range

c (number = 2) Value subtracted from threshold

Instance Members

▸ predict(x)

lib/model/adaptive_thresholding.js

Returns thresholded values.

predict(x: Array<Array<number>>): Array<Array<(0 | 1)>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<(0 | 1)>>: Predicted values

AffinityPropagation

Affinity propagation model

new AffinityPropagation()

Instance Members

▸ categories

Category list

categories

Type: Array<number>

▸ centroids

Centroids

centroids

Type: Array<Array<number>>

▸ size

Number of clusters

size

Type: number

▸ epoch

Learning epoch

epoch

Type: number

▸ init(datas)

Initialize this model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ fit()

Fit this model once.

fit()

▸ predict()

Returns categories corresponding the data.

predict(): Array<number>

Returns

Array<number>: Predicted values

AgglomerativeClusterNode

AgglomerativeClusterNode

Type: object

Properties

point (Array<number>?) : Data point of leaf node

index (number?) : Data index of leaf node

distance (number?) : Distance between children nodes

distances (Array<number>?) : Distances of leaf data and others

size (number) : Number of leaf nodes

children (Array<AgglomerativeClusterNode>?) : Children nodes

leafs (Array<AgglomerativeClusterNode>) : Leaf nodes

AgglomerativeClustering

Agglomerative clustering

new AgglomerativeClustering(metric: ("euclid" | "manhattan" | "chebyshev" | function (Array<number>, Array<number>): number))

Parameters

metric (("euclid" | "manhattan" | "chebyshev" | function (Array<number>, Array<number>): number) = 'euclid') Metric name

Instance Members

▸ fit(points)

Fit model parameters.

fit(points: Array<Array<number>>)

Parameters

points (Array<Array<number>>) Training data

▸ getClusters(number)

Returns the specified number of clusters.

getClusters(number: number): Array<AgglomerativeClusterNode>

Parameters

number (number) Number of clusters

Returns

Array<AgglomerativeClusterNode>: Cluster nodes

▸ predict(k)

Returns predicted categories.

predict(k: number): Array<number>

Parameters

k (number) Number of clusters

Returns

Array<number>: Predicted values

▸ distance(c1, c2)

Returns a distance between two nodes.

distance(c1: AgglomerativeClusterNode, c2: AgglomerativeClusterNode): number

Parameters

c1 (AgglomerativeClusterNode) Node

c2 (AgglomerativeClusterNode) Node

Returns

number: Distance

▸ update(ca, cb, ck, ka, kb, ab)

Returns new distance.

update(ca: number, cb: number, ck: number, ka: number, kb: number, ab: number): number

Parameters

ca (number) Number of datas in a merging node A

cb (number) Number of datas in a merging node B

ck (number) Number of datas in a current node

ka (number) Distance between node A and current node

kb (number) Distance between node B and current node

ab (number) Distance between node A and node B

Returns

number: New distance between current node and merged node

CompleteLinkageAgglomerativeClustering

Complete linkage agglomerative clustering

new CompleteLinkageAgglomerativeClustering()

Instance Members

▸ distance(c1, c2)

Returns a distance between two nodes.

distance(c1: AgglomerativeClusterNode, c2: AgglomerativeClusterNode): number

Parameters

c1 (AgglomerativeClusterNode) Node

c2 (AgglomerativeClusterNode) Node

Returns

number: Distance

▸ update(ca, cb, ck, ka, kb, ab)

Returns new distance.

update(ca: number, cb: number, ck: number, ka: number, kb: number, ab: number): number

Parameters

ca (number) Number of datas in a merging node A

cb (number) Number of datas in a merging node B

ck (number) Number of datas in a current node

ka (number) Distance between node A and current node

kb (number) Distance between node B and current node

ab (number) Distance between node A and node B

Returns

number: New distance between current node and merged node

SingleLinkageAgglomerativeClustering

Single linkage agglomerative clustering

new SingleLinkageAgglomerativeClustering()

Instance Members

▸ distance(c1, c2)

Returns a distance between two nodes.

distance(c1: AgglomerativeClusterNode, c2: AgglomerativeClusterNode): number

Parameters

c1 (AgglomerativeClusterNode) Node

c2 (AgglomerativeClusterNode) Node

Returns

number: Distance

▸ update(ca, cb, ck, ka, kb, ab)

Returns new distance.

update(ca: number, cb: number, ck: number, ka: number, kb: number, ab: number): number

Parameters

ca (number) Number of datas in a merging node A

cb (number) Number of datas in a merging node B

ck (number) Number of datas in a current node

ka (number) Distance between node A and current node

kb (number) Distance between node B and current node

ab (number) Distance between node A and node B

Returns

number: New distance between current node and merged node

GroupAverageAgglomerativeClustering

Group average agglomerative clustering

new GroupAverageAgglomerativeClustering()

Instance Members

▸ distance(c1, c2)

Returns a distance between two nodes.

distance(c1: AgglomerativeClusterNode, c2: AgglomerativeClusterNode): number

Parameters

c1 (AgglomerativeClusterNode) Node

c2 (AgglomerativeClusterNode) Node

Returns

number: Distance

▸ update(ca, cb, ck, ka, kb, ab)

Returns new distance.

update(ca: number, cb: number, ck: number, ka: number, kb: number, ab: number): number

Parameters

ca (number) Number of datas in a merging node A

cb (number) Number of datas in a merging node B

ck (number) Number of datas in a current node

ka (number) Distance between node A and current node

kb (number) Distance between node B and current node

ab (number) Distance between node A and node B

Returns

number: New distance between current node and merged node

WardsAgglomerativeClustering

Ward's agglomerative clustering

new WardsAgglomerativeClustering()

Instance Members

▸ distance(c1, c2)

Returns a distance between two nodes.

distance(c1: AgglomerativeClusterNode, c2: AgglomerativeClusterNode): number

Parameters

c1 (AgglomerativeClusterNode) Node

c2 (AgglomerativeClusterNode) Node

Returns

number: Distance

▸ update(ca, cb, ck, ka, kb, ab)

Returns new distance.

update(ca: number, cb: number, ck: number, ka: number, kb: number, ab: number): number

Parameters

ca (number) Number of datas in a merging node A

cb (number) Number of datas in a merging node B

ck (number) Number of datas in a current node

ka (number) Distance between node A and current node

kb (number) Distance between node B and current node

ab (number) Distance between node A and node B

Returns

number: New distance between current node and merged node

CentroidAgglomerativeClustering

Centroid agglomerative clustering

new CentroidAgglomerativeClustering()

Instance Members

▸ distance(c1, c2)

Returns a distance between two nodes.

distance(c1: AgglomerativeClusterNode, c2: AgglomerativeClusterNode): number

Parameters

c1 (AgglomerativeClusterNode) Node

c2 (AgglomerativeClusterNode) Node

Returns

number: Distance

▸ update(ca, cb, ck, ka, kb, ab)

Returns new distance.

update(ca: number, cb: number, ck: number, ka: number, kb: number, ab: number): number

Parameters

ca (number) Number of datas in a merging node A

cb (number) Number of datas in a merging node B

ck (number) Number of datas in a current node

ka (number) Distance between node A and current node

kb (number) Distance between node B and current node

ab (number) Distance between node A and node B

Returns

number: New distance between current node and merged node

WeightedAverageAgglomerativeClustering

Weighted average agglomerative clustering

new WeightedAverageAgglomerativeClustering()

Instance Members

▸ distance(c1, c2)

Returns a distance between two nodes.

distance(c1: AgglomerativeClusterNode, c2: AgglomerativeClusterNode): number

Parameters

c1 (AgglomerativeClusterNode) Node

c2 (AgglomerativeClusterNode) Node

Returns

number: Distance

▸ update(ca, cb, ck, ka, kb, ab)

Returns new distance.

update(ca: number, cb: number, ck: number, ka: number, kb: number, ab: number): number

Parameters

ca (number) Number of datas in a merging node A

cb (number) Number of datas in a merging node B

ck (number) Number of datas in a current node

ka (number) Distance between node A and current node

kb (number) Distance between node B and current node

ab (number) Distance between node A and node B

Returns

number: New distance between current node and merged node

MedianAgglomerativeClustering

Median agglomerative clustering

new MedianAgglomerativeClustering()

Instance Members

▸ distance(c1, c2)

Returns a distance between two nodes.

distance(c1: AgglomerativeClusterNode, c2: AgglomerativeClusterNode): number

Parameters

c1 (AgglomerativeClusterNode) Node

c2 (AgglomerativeClusterNode) Node

Returns

number: Distance

▸ update(ca, cb, ck, ka, kb, ab)

Returns new distance.

update(ca: number, cb: number, ck: number, ka: number, kb: number, ab: number): number

Parameters

ca (number) Number of datas in a merging node A

cb (number) Number of datas in a merging node B

ck (number) Number of datas in a current node

ka (number) Distance between node A and current node

kb (number) Distance between node B and current node

ab (number) Distance between node A and node B

Returns

number: New distance between current node and merged node

AkimaInterpolation

lib/model/akima.js

Akima interpolation

new AkimaInterpolation(modified: boolean)

Parameters

modified (boolean = false) Use modified method or not

Instance Members

▸ fit(x, y)

lib/model/akima.js

Fit model parameters.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(target)

lib/model/akima.js

Returns predicted interpolated values.

predict(target: Array<number>): Array<number>

Parameters

target (Array<number>) Sample data

Returns

Array<number>: Predicted values

ALMA

Approximate Large Margin algorithm

new ALMA(p: number, alpha: number, b: number, c: number)

Parameters

p (number = 2) Power parameter for norm

alpha (number = 1) Degree of approximation to the optimal margin hyperplane

b (number = 1) Tuning parameter

c (number = 1) Tuning parameter

Instance Members

▸ update(x, y)

Update model parameters with one data.

update(x: Array<number>, y: (1 | -1))

Parameters

x (Array<number>) Training data

y ((1 | -1)) Target value

▸ fit(x, y)

Fit model parameters.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

Returns predicted datas.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

AODE

lib/model/aode.js

Averaged One-Dependence Estimators

new AODE(discrete: number?)

Parameters

discrete (number? = 20) Discretized number

Instance Members

▸ fit(datas, y)

lib/model/aode.js

Fit model.

fit(datas: Array<Array<number>>, y: Array<any>)

Parameters

datas (Array<Array<number>>) Training data

y (Array<any>) Target values

▸ predict(data)

lib/model/aode.js

Returns predicted datas.

predict(data: Array<Array<number>>): Array<(any | null)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(any | null)>: Predicted values

AR

lib/model/ar.js

Autoregressive model

new AR(p: number, method: ("lsm" | "yuleWalker" | "levinson" | "householder")?)

Parameters

p (number) Order

method (("lsm" | "yuleWalker" | "levinson" | "householder")? = 'lsm') Method name

Instance Members

▸ fit(data)

lib/model/ar.js

Fit model.

fit(data: Array<number>)

Parameters

data (Array<number>) Training data

▸ predict(data, k)

lib/model/ar.js

Returns predicted future values.

predict(data: Array<number>, k: number): Array<number>

Parameters

data (Array<number>) Sample data

k (number) Prediction count

Returns

Array<number>: Predicted values

ARMA

lib/model/arma.js

Autoregressive moving average model

new ARMA(p: number, q: number)

Parameters

p (number) Order of AR

q (number) Order of MA

Instance Members

▸ fit(data)

lib/model/arma.js

Fit model.

fit(data: Array<number>)

Parameters

data (Array<number>) Training data

▸ predict(data, k)

lib/model/arma.js

Returns predicted future values.

predict(data: Array<number>, k: number): Array<number>

Parameters

data (Array<number>) Sample data

k (number) Prediction count

Returns

Array<number>: Predicted values

AROW

Adaptive regularization of Weight Vectors

new AROW(r: number)

Parameters

r (number = 0.1) Learning rate

Instance Members

▸ init(train_x, train_y)

Initialize this model.

init(train_x: Array<Array<number>>, train_y: Array<(1 | -1)>)

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<(1 | -1)>) Target values

▸ update(x, y)

Update model parameters with one data.

update(x: Matrix, y: (1 | -1))

Parameters

x (Matrix) Training data

y ((1 | -1)) Target value

▸ fit()

Fit model parameters.

fit()

▸ predict(data)

Returns predicted datas.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

ART

Adaptive resonance theory

new ART(t: number?, method: "l2"?)

Parameters

t (number? = 1) Threshold

method ("l2"? = 'l2') Method name

Instance Members

▸ size

Number of clusters

size

Type: number

▸ fit(datas)

Fit model and returns predicted categories.

fit(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Predicted values

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

Apriori

Apriori algorithm

new Apriori(minsup: number)

Parameters

minsup (number) Minimum support

Instance Members

▸ predict(x)

Returns predicted sets.

predict(x: Array<Array<any>>): Array<SetKeyMap>

Parameters

x (Array<Array<any>>) Training data

Returns

Array<SetKeyMap>: Predicted values

AssociationAnalysis

Association analysis

new AssociationAnalysis(support: number)

Parameters

support (number) Minimum support

Instance Members

▸ fit(x)

Fit model.

fit(x: Array<Array<any>>)

Parameters

x (Array<Array<any>>) Training data

▸ items(n)

Returns appearing keys.

items(n: number): Iterator<Array<string>>

Parameters

n (number = 1) Length of key

Returns

Iterator<Array<string>>: Appearing keys

▸ support(a)

Returns support value.

support(a: ...any): number

Parameters

a (...any) Keys

Returns

number: Support value

▸ confidence(a, b)

Returns confidence value.

confidence(a: any, b: any): number

Parameters

a (any) Key

b (any) Key

Returns

number: Confidence value

▸ lift(a, b)

Returns lift value.

lift(a: any, b: any): number

Parameters

a (any) Key

b (any) Key

Returns

number: Lift value

Autoencoder

Autoencoder

new Autoencoder(input_size: number, reduce_size: number, enc_layers: Array<LayerObject>, dec_layers: Array<LayerObject>, optimizer: string)

Parameters

input_size (number) Input size

reduce_size (number) Reduced dimension

enc_layers (Array<LayerObject>) Layers of encoder

dec_layers (Array<LayerObject>) Layers of decoder

optimizer (string) Optimizer of the network

Instance Members

▸ epoch

Epoch

epoch

Type: number

▸ fit(train_x, iteration, rate, batch, rho)

Fit model.

fit(train_x: Array<Array<number>>, iteration: number, rate: number, batch: number, rho: number): number

Parameters

train_x (Array<Array<number>>) Training data

iteration (number) Iteration count

rate (number) Learning rate

batch (number) Batch size

rho (number) Sparsity parameter

Returns

number: Loss value

▸ predict(x)

Returns predicted datas.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

▸ reduce(x)

lib/model/automatic_thresholding.js

Returns reduced datas.

reduce(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

AutomaticThresholding

Automatic thresholding

new AutomaticThresholding()

Instance Members

▸ fit(x)

lib/model/automatic_thresholding.js

Fit model.

fit(x: Array<number>)

Parameters

x (Array<number>) Training data

▸ predict(x)

lib/model/automatic_thresholding.js

Returns thresholded values.

predict(x: Array<number>): Array<(0 | 1)>

Parameters

x (Array<number>) Sample data

Returns

Array<(0 | 1)>: Predicted values

AverageShiftedHistogram

lib/model/average_shifted_histogram.js

Average shifted histogram

new AverageShiftedHistogram(config: object, step: number)

Parameters

config (object) Config

step (number) Number of bins to average

Instance Members

▸ fit(datas)

lib/model/average_shifted_histogram.js

Returns predicted values.

fit(datas: Array<Array<number>>): Array<any>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<any>: An array nested by the number of dimensions of the data

▸ predict(datas)

lib/model/average_shifted_histogram.js

Returns predicted counted values.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

BalancedHistogramThresholding

lib/model/balanced_histogram.js

Balanced histogram thresholding

new BalancedHistogramThresholding(minCount: number?)

Parameters

minCount (number? = 500) Minimum data count

Instance Members

▸ predict(x)

lib/model/balanced_histogram.js

Returns thresholded values.

predict(x: Array<number>): Array<(0 | 1)>

Parameters

x (Array<number>) Training data

Returns

Array<(0 | 1)>: Predicted values

Ballseptron

lib/model/ballseptron.js

Ballseptron

new Ballseptron(r: number)

Parameters

r (number) Radius

Instance Members

▸ fit(x, y)

lib/model/ballseptron.js

Fit model parameters.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

lib/model/ballseptron.js

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

Banditron

lib/model/banditron.js

Banditron

new Banditron(gamma: number?)

Parameters

gamma (number? = 0.5) Gamma

Instance Members

▸ fit(x, y)

lib/model/banditron.js

Fit model parameters.

fit(x: Array<Array<number>>, y: Array<any>)

Parameters

x (Array<Array<number>>) Training data

y (Array<any>) Target values

▸ predict(data)

lib/model/banditron.js

Returns predicted values.

predict(data: Array<Array<number>>): Array<any>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

BayesianLinearRegression

lib/model/bayesian_linear.js

Bayesian linear regression

new BayesianLinearRegression(lambda: number?, sigma: number?)

Parameters

lambda (number? = 0.1) Tuning parameter

sigma (number? = 0.2) Initial sigma of normal distribution

Instance Members

▸ fit(x, y)

lib/model/bayesian_linear.js

Fit model once.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

lib/model/bayesian_linear.js

Returns predicted datas.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

BayesianNetwork

lib/model/bayesian_network.js

Bayesian Network

new BayesianNetwork(alpha: number)

Parameters

alpha (number) Equivalent sample size

Instance Members

▸ fit(x)

lib/model/bayesian_network.js

Fit model.

fit(x: Array<Array<any>>)

Parameters

x (Array<Array<any>>) Training data

▸ probability(x)

lib/model/bayesian_network.js

Returns probability values.

probability(x: Array<Array<any>>): Array<number>

Parameters

x (Array<Array<any>>) Sample data

Returns

Array<number>: Predicted values

BernsenThresholding

lib/model/bernsen.js

Bernsen thresholding

new BernsenThresholding(n: number?, ct: number?)

Parameters

n (number? = 3) Size of local range

ct (number? = 15) Minimum value of contrast

Instance Members

▸ predict(x)

lib/model/bernsen.js

Returns thresholded values.

predict(x: Array<Array<number>>): Array<Array<(0 | 1)>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<(0 | 1)>>: Predicted values

BesselFilter

lib/model/bessel.js

Bessel filter

new BesselFilter(n: number?, c: number?)

Parameters

n (number? = 2) Order

c (number? = 0.5) Cutoff rate

Instance Members

▸ predict(x)

lib/model/bessel.js

Returns predicted datas.

predict(x: Array<number>): Array<number>

Parameters

x (Array<number>) Training data

Returns

Array<number>: Predicted values

BilinearInterpolation

lib/model/bilinear_interpolation.js

Bilinear interpolation

new BilinearInterpolation()

Instance Members

▸ fit(values, grids)

lib/model/bilinear_interpolation.js

Fit model parameters.

fit(values: Array<Array<number>>, grids: [Array<number>, Array<number>])

Parameters

values (Array<Array<number>>) Training data

grids ([Array<number>, Array<number>]) Grid values for each axises

▸ predict(x)

lib/model/bilinear_interpolation.js

Returns predicted interpolated values.

predict(x: Array<Array<number>>): Array<(number | null)>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<(number | null)>: Predicted values

BIRCH

lib/model/birch.js

Balanced iterative reducing and clustering using hierarchies

new BIRCH(k: number, b: number?, t: number?, l: number?)

Parameters

k (number)

b (number? = 10) Maximum number of entries for each non-leaf nodes

t (number? = 0.2) Threshold

l (number? = Infinity) Maximum number of entries for each leaf nodes

Instance Members

▸ fit(datas)

lib/model/birch.js

Fit model.

fit(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ predict(datas)

lib/model/birch.js

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

BisectingKMeans

Bisecting k-Means algorithm

new BisectingKMeans()

Instance Members

▸ centroids

Centroids

centroids

Type: Array<Array<number>>

▸ fit(datas)

Fit model.

fit(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ predict(datas)

lib/model/brahmagupta_interpolation.js

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

BOGD

lib/model/bogd.js

Bounded Online Gradient Descent

new BOGD(b: number?, eta: number?, lambda: number?, gamma: number?, sampling: ("uniform" | "nonuniform")?, kernel: any, loss: ("zero_one" | "hinge")?)

Parameters

b (number? = 10) Maximum budget size

eta (number? = 1) Stepsize

lambda (number? = 0.1) Regularization parameter

gamma (number? = 0.1) Maximum coefficient

sampling (("uniform" | "nonuniform")? = 'nonuniform') Sampling approach

kernel (any = 'gaussian')

loss (("zero_one" | "hinge")? = 'hinge') Loss type name

Instance Members

▸ fit(x, y)

lib/model/bogd.js

Fit model.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

lib/model/bogd.js

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

BoxCox

lib/model/box_cox.js

Box-Cox transformation

new BoxCox(lambda: number?)

Parameters

lambda (number? = null) Lambda

Instance Members

▸ fit(x)

lib/model/box_cox.js

Fit model.

fit(x: (Array<number> | Array<Array<number>>))

Parameters

x ((Array<number> | Array<Array<number>>)) Training data

▸ predict(x)

lib/model/box_cox.js

Returns transformed values.

predict(x: (Array<number> | Array<Array<number>>)): (Array<number> | Array<Array<number>>)

Parameters

x ((Array<number> | Array<Array<number>>)) Sample data

Returns

(Array<number> | Array<Array<number>>): Predicted values

BPA

lib/model/bpa.js

Budgeted online Passive-Aggressive

new BPA(c: number?, b: number?, version: ("simple" | "projecting" | "nn")?, kernel: any)

Parameters

c (number? = 1) Regularization parameter

b (number? = 10) Budget size

version (("simple" | "projecting" | "nn")? = 'simple') Version

kernel (any = 'gaussian')

Instance Members

▸ fit(x, y)

lib/model/bpa.js

Fit model.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

lib/model/bpa.js

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

BrahmaguptaInterpolation

Brahmagupta interpolation

new BrahmaguptaInterpolation()

Instance Members

▸ fit(x, y)

lib/model/brahmagupta_interpolation.js

Fit model parameters.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(target)

lib/model/brahmagupta_interpolation.js

Returns predicted interpolated values.

predict(target: Array<number>): Array<number>

Parameters

target (Array<number>) Sample data

Returns

Array<number>: Predicted values

BRIDGE

lib/model/bridge.js

BRIDGE

new BRIDGE(k: number, e_core: number, e_den: number)

Parameters

k (number) K-means clustering size

e_core (number) e for core distance

e_den (number) e for density base clustering

Instance Members

▸ predict(datas)

lib/model/bridge.js

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Predicted values

BSGD

Budgeted Stochastic Gradient Descent

new BSGD(b: number?, eta: number?, lambda: number?, maintenance: ("removal" | "projection" | "merging")?, kernel: any)

Parameters

b (number? = 10) Budget size

eta (number? = 1) Learning rate

lambda (number? = 1) Regularization parameter

maintenance (("removal" | "projection" | "merging")? = 'removal') Maintenance type

kernel (any = 'gaussian')

Instance Members

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

MulticlassBSGD

Multiclass Budgeted Stochastic Gradient Descent

new MulticlassBSGD(b: number?, eta: number?, lambda: number?, maintenance: ("removal" | "projection" | "merging")?, kernel: any)

Parameters

b (number? = 10) Budget size

eta (number? = 1) Learning rate

lambda (number? = 1) Regularization parameter

maintenance (("removal" | "projection" | "merging")? = 'removal') Maintenance type

kernel (any = 'gaussian')

Instance Members

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<any>)

Parameters

x (Array<Array<number>>) Training data

y (Array<any>) Target values

▸ predict(data)

lib/model/budget_perceptron.js

Returns predicted values.

predict(data: Array<Array<number>>): Array<any>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

BudgetPerceptron

Budget Perceptron

new BudgetPerceptron(beta: number, n: number?)

Parameters

beta (number) Tolerance

n (number?) Cachs size

Instance Members

▸ fit(x, y)

lib/model/budget_perceptron.js

Fit model parameters.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

lib/model/budget_perceptron.js

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

ButterworthFilter

lib/model/butterworth.js

Butterworth filter

new ButterworthFilter(n: number?, c: number?)

Parameters

n (number? = 2) Order

c (number? = 0.5) Cutoff rate

Instance Members

▸ predict(x)

lib/model/butterworth.js

Returns predicted datas.

predict(x: Array<number>): Array<number>

Parameters

x (Array<number>) Training data

Returns

Array<number>: Predicted values

C2P

Clustering based on Closest Pairs

new C2P(r: number, m: number)

Parameters

r (number) Number of representative points

m (number) Number of required sub-clusters

Instance Members

▸ fit(data)

Fit model.

fit(data: Array<Array<number>>)

Parameters

data (Array<Array<number>>) Training data

▸ getClusters(number)

Returns the specified number of clusters.

getClusters(number: number)

Parameters

number (number) Number of clusters

▸ predict(k)

Returns predicted categories.

predict(k: number): Array<number>

Parameters

k (number) Number of clusters

Returns

Array<number>: Predicted values

Canny

lib/model/canny.js

Canny edge detection

new Canny(th1: number, th2: number)

Parameters

th1 (number) Big threshold

th2 (number) Small threshold

Instance Members

▸ predict(x)

lib/model/canny.js

Returns predicted edge flags.

predict(x: Array<Array<number>>): Array<Array<boolean>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<boolean>>: Predicted values. true if a pixel is edge.

CAST

Clustering Affinity Search Technique

new CAST(t: number)

Parameters

t (number) Affinity threshold

Instance Members

▸ size

Number of clusters

size

Type: number

▸ fit(datas)

Fit model.

fit(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ predict()

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

CategoricalNaiveBayes

Categorical naive bayes

new CategoricalNaiveBayes(alpha: number?)

Parameters

alpha (number? = 1.0) Smoothing parameter

Instance Members

▸ fit(datas, labels)

Fit model.

fit(datas: Array<Array<any>>, labels: Array<any>)

Parameters

datas (Array<Array<any>>) Training data

labels (Array<any>) Target values

▸ probability(datas)

Returns predicted probabilities.

probability(datas: Array<Array<any>>): Array<Array<number>>

Parameters

datas (Array<Array<any>>) Sample data

Returns

Array<Array<number>>: Predicted values

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<any>>): Array<any>

Parameters

datas (Array<Array<any>>) Sample data

Returns

Array<any>: Predicted values

CatmullRomSplines

Catmull-Rom splines interpolation

new CatmullRomSplines()

Instance Members

▸ fit(x, y)

Fit model parameters.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(target)

Returns predicted interpolated values.

predict(target: Array<number>): Array<number>

Parameters

target (Array<number>) Sample data

Returns

Array<number>: Predicted values

CentripetalCatmullRomSplines

Centripetal Catmull-Rom splines interpolation

new CentripetalCatmullRomSplines(alpha: number)

Parameters

alpha (number = 0.5) Number for knot parameterization

Instance Members

▸ fit(x, y)

Fit model parameters.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(target)

Returns predicted interpolated values.

predict(target: Array<number>): Array<number>

Parameters

target (Array<number>) Sample data

Returns

Array<number>: Predicted values

CHAMELEON

CHAMELEON

new CHAMELEON(k: number?)

Parameters

k (number? = 5) Number of neighborhoods

Instance Members

▸ fit(datas)

Fit model.

fit(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ getClusters(number)

Returns the specified number of clusters.

getClusters(number: number): Array<any>

Parameters

number (number) Number of clusters

Returns

Array<any>: Cluster nodes

▸ predict(k)

lib/model/change_finder.js

Returns predicted categories.

predict(k: number): Array<number>

Parameters

k (number) Number of clusters

Returns

Array<number>: Predicted values

ChangeFinder

Change finder

new ChangeFinder(p: number?, r: number?, smooth: number?)

Deprecated: Does not work properly

Parameters

p (number? = 1) Order

r (number? = 0.5) Forgetting factor

smooth (number? = 10) Smoothing window size

Instance Members

▸ fit(datas)

lib/model/change_finder.js

Fit model.

fit(datas: Array<number>)

Parameters

datas (Array<number>) Training data

▸ predict()

lib/model/change_finder.js

Returns predicted scores.

predict(): Array<number>

Returns

Array<number>: Predicted values

ChebyshevFilter

lib/model/chebyshev.js

Chebyshev filter

new ChebyshevFilter(type: (1 | 2)?, ripple: number?, n: number?, c: number?)

Parameters

type ((1 | 2)? = 1) Type number

ripple (number? = 1) Ripple factor

n (number? = 2) Order

c (number? = 0.5) Cutoff rate

Instance Members

▸ predict(x)

lib/model/chebyshev.js

Returns predicted datas.

predict(x: Array<number>): Array<number>

Parameters

x (Array<number>) Training data

Returns

Array<number>: Predicted values

CLARA

Clustering LARge Applications

new CLARA(k: number)

Parameters

k (number) Number of clusters

Instance Members

▸ init(datas)

Initialize model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ fit()

Fit model.

fit()

▸ predict()

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

CLARANS

Clustering Large Applications based on RANdomized Search

new CLARANS(k: number)

Parameters

k (number) Number of clusters

Instance Members

▸ init(datas)

Initialize model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ fit(numlocal, maxneighbor)

Fit model once.

fit(numlocal: number, maxneighbor: number)

Parameters

numlocal (number) Iteration count for local

maxneighbor (number) Iteration count for neighborhoods

▸ predict()

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

CLIQUE

CLustering In QUEst

new CLIQUE(xi: (number | Array<number>), t: number)

Parameters

xi ((number | Array<number>)) Intervals

t (number) Density threshold

Instance Members

▸ size

Number of clusters

size

Type: number

▸ fit(datas)

Fit model.

fit(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ predict(x)

Returns predicted categories.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted categories

CLUES

CLUstEring based on local Shrinking

new CLUES(alpha: number?)

Parameters

alpha (number? = 0.05) Speed factor

Instance Members

▸ size

Number of clusters

size

Type: number

▸ fit(datas)

Fit model.

fit(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ predict()

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

CoTraining

Co-training

new CoTraining(view1: object, view2: object)

Parameters

view1 (object) View

view2 (object) View

Instance Members

▸ init(x, y)

Initialize model.

init(x: Array<Array<number>>, y: Array<(any | null)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(any | null)>) Target values

▸ fit()

Fit model.

fit()

▸ predict()

Returns predicted categories.

predict(): Array<(any | null)>

Returns

Array<(any | null)>: Predicted values

COF

lib/model/cof.js

Connectivity-based Outlier Factor

new COF(k: number)

Parameters

k (number) Number of neighborhoods

Instance Members

▸ predict(datas)

lib/model/cof.js

Returns anomaly degrees.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Predicted values

COLL

Conscience on-line learning

new COLL(c: number, eta: number?, kernel: any)

Parameters

c (number) Number of clusters

eta (number? = 1) Initial learning rate

kernel (any = 'gaussian')

Instance Members

▸ init(datas)

Initialize model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ fit()

Fit model once.

fit(): number

Returns

number: Convergence criterion

▸ predict()

lib/model/complement_naive_bayes.js

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

ComplementNaiveBayes

Complement Naive Bayes

new ComplementNaiveBayes(distribution: "gaussian"?)

Parameters

distribution ("gaussian"? = 'gaussian') Distribution name

Instance Members

▸ fit(datas, labels)

lib/model/complement_naive_bayes.js

Fit model.

fit(datas: Array<Array<number>>, labels: Array<any>)

Parameters

datas (Array<Array<number>>) Training data

labels (Array<any>) Target values

▸ predict(data)

lib/model/complement_naive_bayes.js

Returns predicted categories.

predict(data: Array<Array<number>>): Array<any>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

ConfidenceWeighted

Confidence weighted

new ConfidenceWeighted(eta: number)

Parameters

eta (number) Confidence value

Instance Members

▸ init(train_x, train_y)

Initialize this model.

init(train_x: Array<Array<number>>, train_y: Array<(1 | -1)>)

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<(1 | -1)>) Target values

▸ update(x, y)

Update model parameters with one data.

update(x: Matrix, y: (1 | -1))

Parameters

x (Matrix) Training data

y ((1 | -1)) Target value

▸ fit()

Fit model parameters.

fit()

▸ predict(data)

Returns predicted datas.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

SoftConfidenceWeighted

lib/model/cosine_interpolation.js

Soft confidence weighted

new SoftConfidenceWeighted(eta: number, cost: number, v: (1 | 2))

Extends ConfidenceWeighted

Parameters

eta (number) Confidence value

cost (number) Tradeoff value between passiveness and aggressiveness

v ((1 | 2)) Version number

CosineInterpolation

Cosine interpolation

new CosineInterpolation()

Instance Members

▸ fit(x, y)

lib/model/cosine_interpolation.js

Fit model parameters.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(target)

lib/model/cosine_interpolation.js

Returns predicted interpolated values.

predict(target: Array<number>): Array<number>

Parameters

target (Array<number>) Sample data

Returns

Array<number>: Predicted values

CRF

Conditional random fields

new CRF()

Instance Members

▸ fit(x, y)

Fit model.

fit(x: Array<Array<any>>, y: Array<Array<any>>)

Parameters

x (Array<Array<any>>) Training data

y (Array<Array<any>>) Target values

▸ probability(x, y)

Returns probability P(y|x).

probability(x: Array<any>, y: Array<any>): number

Parameters

x (Array<any>) Sample data

y (Array<any>) Target values

Returns

number: Predicted values

▸ predict(x)

lib/model/cubic_convolution.js

Returns predicted labels.

predict(x: Array<Array<any>>): Array<Array<any>>

Parameters

x (Array<Array<any>>) Sample data

Returns

Array<Array<any>>: Predicted values

CubicConvolutionInterpolation

Cubic-convolution interpolation

new CubicConvolutionInterpolation(a: number)

Parameters

a (number) Tuning parameter

Instance Members

▸ fit(values)

lib/model/cubic_convolution.js

Fit model parameters.

fit(values: Array<number>)

Parameters

values (Array<number>) Training data

▸ predict(index)

lib/model/cubic_convolution.js

Returns predicted interpolated values.

predict(index: Array<number>): Array<number>

Parameters

index (Array<number>) Sample data

Returns

Array<number>: Predicted values

CubicHermiteSpline

lib/model/cubic_hermite_spline.js

Cubic Hermite spline

new CubicHermiteSpline(t: number, b: number)

Parameters

t (number) Tension factor

b (number) Bias factor

Instance Members

▸ fit(x, y)

lib/model/cubic_hermite_spline.js

Fit model parameters.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(target)

lib/model/cubic_hermite_spline.js

Returns predicted interpolated values.

predict(target: Array<number>): Array<number>

Parameters

target (Array<number>) Sample data

Returns

Array<number>: Predicted values

CubicInterpolation

lib/model/cubic_interpolation.js

Cubic interpolation

new CubicInterpolation()

Instance Members

▸ fit(x, y)

lib/model/cubic_interpolation.js

Fit model parameters.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(target)

lib/model/cubic_interpolation.js

Returns predicted interpolated values.

predict(target: Array<number>): Array<number>

Parameters

target (Array<number>) Target values

Returns

Array<number>: Predicted values

CumulativeMovingAverage

lib/model/cumulative_moving_average.js

Cumulative moving average

new CumulativeMovingAverage()

Instance Members

▸ predict(data)

lib/model/cumulative_moving_average.js

Returns smoothed values.

predict(data: Array<number>): Array<number>

Parameters

data (Array<number>) Training data

Returns

Array<number>: Predicted values

CumSum

Cumulative sum change point detection

new CumSum()

Instance Members

▸ init(datas)

Initialize model.

init(datas: Array<number>)

Parameters

datas (Array<number>) Training data

▸ fit()

Fit model once.

fit()

▸ predict()

Returns predicted values.

predict(): Array<boolean>

Returns

Array<boolean>: Predicted values

CURENode

CURENode

Type: object

Properties

point (Array<number>?) : Data point of leaf node

index (number?) : Data index of leaf node

distance (number?) : Distance between children nodes

size (number) : Number of leaf nodes

children (Array<CURENode>?) : Children nodes

leafs (Array<CURENode>) : Leaf nodes

CURE

Clustering Using REpresentatives

new CURE(c: number)

Parameters

c (number) Number of representative points

Instance Members

▸ fit(data)

Fit model.

fit(data: Array<Array<number>>)

Parameters

data (Array<Array<number>>) Training data

▸ getClusters(number)

Returns the specified number of clusters.

getClusters(number: number): Array<CURENode>

Parameters

number (number) Number of clusters

Returns

Array<CURENode>: Cluster nodes

▸ predict(k)

Returns predicted categories.

predict(k: number): Array<number>

Parameters

k (number) Number of clusters

Returns

Array<number>: Predicted values

DiscriminantAdaptiveNearestNeighbor

lib/model/dann.js

Discriminant adaptive nearest neighbor

new DiscriminantAdaptiveNearestNeighbor(k: number?, iteration: number?)

Parameters

k (number? = null) Number of neighborhoods

iteration (number? = 1) Iteration count

Instance Members

▸ fit(x, y)

lib/model/dann.js

Fit model.

fit(x: Array<Array<number>>, y: Array<any>)

Parameters

x (Array<Array<number>>) Training data

y (Array<any>) Target values

▸ predict(data)

lib/model/dann.js

Returns predicted categories.

predict(data: Array<Array<number>>): Array<any>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

DBCLASD

lib/model/dbclasd.js

Distribution Based Clustering of LArge Spatial Databases

new DBCLASD()

Instance Members

▸ predict(datas)

lib/model/dbclasd.js

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Predicted values

DBSCAN

lib/model/dbscan.js

Density-based spatial clustering of applications with noise

new DBSCAN(eps: number?, minPts: number?, metric: ("euclid" | "manhattan" | "chebyshev" | function (Array<number>, Array<number>): number)?)

Parameters

eps (number? = 0.5) Radius to determine neighborhood

minPts (number? = 5) Minimum size of cluster

metric (("euclid" | "manhattan" | "chebyshev" | function (Array<number>, Array<number>): number)? = 'euclid') Metric name

Instance Members

▸ predict(datas)

lib/model/dbscan.js

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Predicted values

DecisionTree

Decision tree

new DecisionTree()

Instance Members

▸ depth

Depth of the tree

depth

Type: number

▸ init(datas, targets)

Initialize model.

init(datas: Array<Array<number>>, targets: Array<any>)

Parameters

datas (Array<Array<number>>) Training data

targets (Array<any>) Target values

▸ fit()

Fit model.

fit()

▸ importance()

Returns importances of the features.

importance(): Array<number>

Returns

Array<number>: Importances

▸ predict_value(data)

Returns predicted values.

predict_value(data: Array<Array<number>>): Array<number>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

DecisionTreeClassifier

Decision tree classifier

new DecisionTreeClassifier(method: ("ID3" | "CART"))

Extends DecisionTree

Parameters

method (("ID3" | "CART")) Method name

Instance Members

▸ predict_prob(data)

Returns probability of the datas.

predict_prob(data: Array<Array<number>>): Array<number>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

▸ predict(data)

Returns predicted values.

predict(data: Array<Array<number>>): Array<any>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

DecisionTreeRegression

Decision tree regression

new DecisionTreeRegression()

Extends DecisionTree

Instance Members

▸ predict(data)

lib/model/delaunay_interpolation.js

Returns predicted values.

predict(data: Array<Array<number>>): Array<number>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

DelaunayInterpolation

Delaunay interpolation

new DelaunayInterpolation()

Instance Members

▸ fit(x, y)

lib/model/delaunay_interpolation.js

Fit model.

fit(x: Array<Array<number>>, y: Array<number>)

Parameters

x (Array<Array<number>>) Training data

y (Array<number>) Target values

▸ predict(x)

lib/model/delaunay_interpolation.js

Returns probabilities of the datas.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

DemingRegression

lib/model/deming.js

Deming regression

new DemingRegression(d: number)

Parameters

d (number) Ratio of variances

Instance Members

▸ fit(x, y)

lib/model/deming.js

Fit model.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(x)

lib/model/deming.js

Returns predicted values.

predict(x: Array<number>): Array<number>

Parameters

x (Array<number>) Sample data

Returns

Array<number>: Predicted values

DENCLUE

DENsity CLUstering

new DENCLUE(h: number, version: (1 | 2)?, kernel: ("gaussian" | {name: "gaussian"} | function (Array<number>): number)?)

Parameters

h (number) Smoothing parameter for the kernel

version ((1 | 2)? = 1) Version number

kernel (("gaussian" | {name: "gaussian"} | function (Array<number>): number)? = 'gaussian') Kernel name

Instance Members

▸ size

Number of clusters

size

Type: number

▸ init(datas)

Initialize model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ fit()

Fit model.

fit()

▸ predict()

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

DIANA

DIvisive ANAlysis Clustering

new DIANA()

Instance Members

▸ size

Number of clusters

size

Type: number

▸ init(datas)

Initialize model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ fit()

Fit model.

fit()

▸ predict()

lib/model/diffusion_map.js

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

DiffusionMap

Diffusion map

new DiffusionMap(t: number, rd: number, kernel: ("gaussian" | {name: "gaussian"} | function (Array<number>, Array<number>): number)?)

Parameters

t (number) Power parameter

rd (number) Reduced dimension

kernel (("gaussian" | {name: "gaussian"} | function (Array<number>, Array<number>): number)? = 'gaussian') Kernel name

Instance Members

▸ predict(x)

lib/model/diffusion_map.js

Returns reduced values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<number>>: Predicted values

DQNAgent

Deep Q-Network agent

new DQNAgent(env: RLEnvironmentBase, resolution: number, layers: Array<LayerObject>, optimizer: string)

Parameters

env (RLEnvironmentBase) Environment

resolution (number) Resolution of actions

layers (Array<LayerObject>) Network layers

optimizer (string) Optimizer of the network

Instance Members

▸ method

DQN Method

method

Parameters

value (("DQN" | "DDQN")) New method name

▸ get_score()

Returns a score.

get_score(): Array<Array<Array<number>>>

Returns

Array<Array<Array<number>>>: Score values

▸ get_action(state, greedy_rate)

Returns a action.

get_action(state: Array<any>, greedy_rate: number): Array<any>

Parameters

state (Array<any>) Current states

greedy_rate (number = 0.002) Greedy rate

Returns

Array<any>: Action

▸ update(action, state, next_state, reward, done, learning_rate, batch)

Update model.

update(action: Array<any>, state: Array<any>, next_state: Array<any>, reward: number, done: boolean, learning_rate: number, batch: number)

Parameters

action (Array<any>) Action

state (Array<any>) Current states

next_state (Array<any>) Next states

reward (number) Reward

done (boolean) Done epoch or not

learning_rate (number) Learning rate

batch (number) Batch size

DrakeKMeans

Drake's accelerated k-Means algorithm

new DrakeKMeans(k: number)

Parameters

k (number) Number of clusters

Instance Members

▸ centroids

Centroids

centroids

Type: Array<Array<number>>

▸ init(datas)

Initialize this model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ fit()

Fit model.

fit()

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

DPAgent

Dynamic programming agent

new DPAgent(env: RLEnvironmentBase, resolution: number?)

Parameters

env (RLEnvironmentBase) Environment

resolution (number? = 20) Resolution

Instance Members

▸ get_score()

Returns a score.

get_score(): Array<Array<Array<number>>>

Returns

Array<Array<Array<number>>>: Score values

▸ get_action(state)

Returns a action.

get_action(state: Array<any>): Array<any>

Parameters

state (Array<any>) Current states

Returns

Array<any>: Action

▸ update(method)

Update model.

update(method: ("value" | "policy"))

Parameters

method (("value" | "policy")) Method name

ElasticNet

Elastic net

new ElasticNet(lambda: number?, alpha: number?, method: ("ISTA" | "CD")?)

Parameters

lambda (number? = 0.1) Regularization strength

alpha (number? = 0.5) Mixing parameter

method (("ISTA" | "CD")? = 'CD') Method name

Instance Members

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

▸ importance()

Returns importances of the features.

importance(): Array<number>

Returns

Array<number>: Importances

ElkanKMeans

Elkan's accelerated k-Means algorithm

new ElkanKMeans(k: number)

Parameters

k (number) Number of clusters

Instance Members

▸ centroids

Centroids

centroids

Type: Array<Array<number>>

▸ init(datas)

Initialize this model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ fit()

Fit model.

fit()

▸ predict(datas)

lib/model/elliptic_filter.js

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

EllipticFilter

Elliptic filter

new EllipticFilter(ripple: number?, n: number?, xi: number?, c: number?)

Parameters

ripple (number? = 1) Ripple factor

n (number? = 2) Order

xi (number? = 1) Selectivity factor

c (number? = 0.5) Cutoff rate

Instance Members

▸ predict(x)

lib/model/elliptic_filter.js

Returns predicted datas.

predict(x: Array<number>): Array<number>

Parameters

x (Array<number>) Training data

Returns

Array<number>: Predicted values

ENaN

lib/model/enan.js

Extended Natural Neighbor

new ENaN(metric: ("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)?)

Parameters

metric (("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)? = 'euclid') Metric name

Instance Members

▸ fit(datas, targets)

lib/model/enan.js

Add datas.

fit(datas: Array<Array<number>>, targets: Array<any>)

Parameters

datas (Array<Array<number>>) Training data

targets (Array<any>) Target values

▸ predict(datas)

lib/model/enan.js

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<any>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

ENN

lib/model/enn.js

Extended Nearest Neighbor

Parameters

version ((0 | 1 | 2)? = 1) Version

k (number? = 5) Number of neighborhoods

metric (("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)? = 'euclid') Metric name

Instance Members

▸ fit(datas, targets)

lib/model/enn.js

Add datas.

fit(datas: Array<Array<number>>, targets: Array<any>)

Parameters

datas (Array<Array<number>>) Training data

targets (Array<any>) Target values

▸ predict(datas)

lib/model/enn.js

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<any>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

BinaryModel

BinaryModel

Type: object

Properties

init (function (Array<Array<number>>, Array<any>): void) : Initialize model

fit (function (...any): void) : Fit model

predict (function (Array<Array<number>>): Array<number>) : Returns predicted values

EnsembleBinaryModel

Ensemble binary models

new EnsembleBinaryModel(model: any, type: ("oneone" | "onerest"), classes: Array<any>?)

Parameters

model (any)

type (("oneone" | "onerest")) Type name

classes (Array<any>?) Initial class labels

Instance Members

▸ init(train_x, train_y)

Initialize model.

init(train_x: Array<Array<number>>, train_y: Array<any>)

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<any>) Target values

▸ fit(x, y, args)

Fit model.

fit(x: Array<Array<number>>, y: Array<any>, args: ...any)

Parameters

x (Array<Array<number>>) Training data

y (Array<any>) Target values

args (...any) Arguments for fit

▸ predict(data)

Returns predicted values.

predict(data: Array<Array<number>>): Array<any>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

ExponentialMovingAverage

Exponential moving average

new ExponentialMovingAverage(k: number)

Parameters

k (number) Degree of weighting decrease

Instance Members

▸ predict(data)

Returns smoothed values.

predict(data: Array<number>): Array<number>

Parameters

data (Array<number>) Training data

Returns

Array<number>: Predicted values

ModifiedMovingAverage

Modified moving average

new ModifiedMovingAverage(k: number)

Parameters

k (number) Degree of weighting decrease

Instance Members

▸ predict(data)

Returns smoothed values.

predict(data: Array<number>): Array<number>

Parameters

data (Array<number>) Training data

Returns

Array<number>: Predicted values

ExtraTrees

Bsae class for Extremely Randomized Trees

new ExtraTrees(tree_num: number, sampling_rate: number?)

Parameters

tree_num (number) Number of trees

sampling_rate (number? = 1.0) Sampling rate

Instance Members

▸ depth

The max depth among the trees.

depth

Type: number

▸ init(datas, targets)

Initialize model.

init(datas: Array<Array<number>>, targets: Array<any>)

Parameters

datas (Array<Array<number>>) Training data

targets (Array<any>) Target values

▸ fit()

Fit model.

fit()

ExtraTreesClassifier

Extra trees classifier

new ExtraTreesClassifier(tree_num: number, sampling_rate: number?)

Extends ExtraTrees

Parameters

tree_num (number) Number of trees

sampling_rate (number? = 1.0) Sampling rate

Instance Members

▸ predict(datas)

Returns predicted values.

predict(datas: Array<Array<number>>): Array<any>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

ExtraTreesRegressor

Extra trees regressor

new ExtraTreesRegressor(tree_num: number, sampling_rate: number?)

Extends ExtraTrees

Parameters

tree_num (number) Number of trees

sampling_rate (number? = 1.0) Sampling rate

Instance Members

▸ predict(datas)

Returns predicted values.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

FastMap

lib/model/fastmap.js

FastMap

new FastMap(rd: number)

Parameters

rd (number) Reduced dimension

Instance Members

▸ predict(x)

lib/model/fastmap.js

Returns reduced values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<number>>: Predicted values

FINDIT

a Fast and INtelligent subspace clustering algorithm using DImension voting

new FINDIT(minsize: number, mindist: number)

Parameters

minsize (number) Mininum size of clusters

mindist (number) Merge threshold

Instance Members

▸ size

Number of clusters

size

Type: number

▸ fit(datas)

Fit model.

fit(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ predict(x)

Returns predicted categories.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted categories

Forgetron

lib/model/forgetron.js

Forgetron

new Forgetron(b: number, kernel: any)

Parameters

b (number) Budget parameter

kernel (any = 'gaussian')

Instance Members

▸ fit(x, y)

lib/model/forgetron.js

Fit model parameters.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

lib/model/forgetron.js

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

FuzzyCMeans

Fuzzy c-means

new FuzzyCMeans(m: number?)

Parameters

m (number? = 2) Fuzziness factor

Instance Members

▸ init(datas)

Initialize model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ add()

Add a new cluster.

add()

▸ fit()

Fit model.

fit()

▸ predict()

Returns predicted coefficients.

predict(): Array<Array<number>>

Returns

Array<Array<number>>: Predicted values

FuzzyKNN

Fuzzy k-nearest neighbor

new FuzzyKNN(k: number?, m: number?)

Parameters

k (number? = 5) Number of neighborhoods

m (number? = 2) Factor of weight for distance

Instance Members

▸ categories

Category list

categories

Type: Array<any>

▸ add(point, category?)

Add a data.

add(point: Array<number>, category: any?)

Parameters

point (Array<number>) Training data

category (any?) Target value

▸ fit(datas, targets)

Add datas.

fit(datas: Array<Array<number>>, targets: Array<any>)

Parameters

datas (Array<Array<number>>) Training data

targets (Array<any>) Target values

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<Array<number>>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

GAN

Generative adversarial networks

new GAN(noise_dim: number, g_hidden: Array<LayerObject>, d_hidden: Array<LayerObject>, g_opt: string, d_opt: string, class_size: (number | null), type: ("" | "conditional"))

Parameters

noise_dim (number) Number of noise dimension

g_hidden (Array<LayerObject>) Layers of generator

d_hidden (Array<LayerObject>) Layers of discriminator

g_opt (string) Optimizer of the generator network

d_opt (string) Optimizer of the discriminator network

class_size ((number | null)) Class size for conditional type

type (("" | "conditional")) Type name

Instance Members

▸ epoch

Epoch

epoch

Type: number

▸ fit(x, y, step, gen_rate, dis_rate, batch)

Fit model.

fit(x: Array<Array<number>>, y: (Array<Array<number>> | null), step: number, gen_rate: number, dis_rate: number, batch: number): {generatorLoss: number, discriminatorLoss: number}

Parameters

x (Array<Array<number>>) Training data

y ((Array<Array<number>> | null)) Conditional values

step (number) Iteration count

gen_rate (number) Learning rate for generator

dis_rate (number) Learning rate for discriminator

batch (number) Batch size

Returns

{generatorLoss: number, discriminatorLoss: number}: Loss value

▸ prob(x, y)

Returns probabilities of the data is true.

prob(x: Array<Array<number>>, y: any): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

y (any) Conditional values

Returns

Array<Array<number>>: Predicted values

▸ generate(n, y)

lib/model/gasser_muller.js

Returns generated data from the model.

generate(n: number, y: (Array<Array<number>> | null)): Array<Array<number>>

Parameters

n (number) Number of generated data

y ((Array<Array<number>> | null)) Conditional values

Returns

Array<Array<number>>: Generated values

GasserMuller

Gasser–Müller kernel estimator

new GasserMuller(h: number)

Parameters

h (number) Smoothing parameter for the kernel

Instance Members

▸ fit(x, y)

lib/model/gasser_muller.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

lib/model/gasser_muller.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

GaussianProcess

Gaussian process

new GaussianProcess(kernel: "gaussian"?, beta: number?)

Parameters

kernel ("gaussian"? = 'gaussian') Kernel name

beta (number? = 1) Precision parameter

Instance Members

▸ init(x, y)

Initialize model.

init(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ fit(learning_rate)

Fit model.

fit(learning_rate: number)

Parameters

learning_rate (number = 0.1) Learning rate

▸ predict(x)

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

GBDT

Gradient boosting decision tree

new GBDT(maxdepth: number?, srate: number?, lr: number?)

Parameters

maxdepth (number? = 1) Maximum depth of tree

srate (number? = 1.0) Sampling rate

lr (number? = 0) Learning rate

Instance Members

▸ size

Number of trees

size

Type: number

▸ init(x, y)

Initialize model.

init(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ fit()

Fit model.

fit()

▸ predict(x)

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

GBDTClassifier

Gradient boosting decision tree classifier

new GBDTClassifier(maxdepth: number?, srate: number?, lr: number?)

Extends GBDT

Parameters

maxdepth (number? = 1) Maximum depth of tree

srate (number? = 1.0) Sampling rate

lr (number? = 0) Learning rate

Instance Members

▸ init(x, y)

Initialize model.

init(x: Array<Array<number>>, y: Array<any>)

Parameters

x (Array<Array<number>>) Training data

y (Array<any>) Target values

▸ predict(x)

lib/model/generalized_esd.js

Returns predicted categories.

predict(x: Array<Array<number>>): Array<any>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

GeneralizedESD

Generalized extreme studentized deviate

new GeneralizedESD(alpha: number, r: number)

Parameters

alpha (number) Significance level

r (number) Max number of outliers

Instance Members

▸ predict(data)

lib/model/generalized_esd.js

Returns a list of the data predicted as outliers or not.

predict(data: Array<Array<number>>): Array<boolean>

Parameters

data (Array<Array<number>>) Training data

Returns

Array<boolean>: Predicted values

GeneticModel

GeneticModel

Type: object

Properties

run (function (...any): void) : Run model

mutation (function (): GeneticModel) : Returns mutated model

mix (function (GeneticModel): GeneticModel) : Returns mixed model

score (function (): number) : Returns a number how good the model is

GeneticAlgorithm

Genetic algorithm

new GeneticAlgorithm(size: number, model: any)

Parameters

size (number) Number of models per generation

model (any)

Instance Members

▸ models

Models

models

Type: Array<GeneticModel>

▸ bestModel

The best model.

bestModel

Returns

GeneticModel: Best model

▸ run(args)

Run for all models.

run(args: ...any)

Parameters

args (...any) Arguments for run

▸ next(mutation_rate = 0.001)

Update models to new generation.

next(mutation_rate: number?)

Parameters

mutation_rate (number? = 0.001) Mutation rate

GeneticAlgorithmGeneration

Genetic algorithm generation

new GeneticAlgorithmGeneration(env: RLEnvironmentBase, size: number?, resolution: number?)

Parameters

env (RLEnvironmentBase) Environment

size (number? = 100) Number of models per generation

resolution (number? = 20) Resolution

Instance Members

▸ reset()

Reset all agents.

reset()

▸ get_score()

Returns a score.

get_score(): Array<Array<Array<number>>>

Returns

Array<Array<Array<number>>>: Score values

▸ top_agent()

Returns the best score agent.

top_agent(): GeneticAlgorithmAgent

Returns

GeneticAlgorithmAgent: Best agent

▸ run()

Run for all agents.

run()

▸ next(mutation_rate)

Update agent to new generation.

next(mutation_rate: number)

Parameters

mutation_rate (number = 0.001) Mutation rate

GeneticKMeans

Genetic k-means model

new GeneticKMeans(k: number, size: number)

Parameters

k (number) Number of clusters

size (number) Number of models per generation

Instance Members

▸ centroids

Centroids

centroids

Type: Array<Array<number>>

▸ bestModel

The best model.

bestModel

Type: GeneticKMeansModel

▸ init(datas)

Initialize model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

▸ fit()

Fit model.

fit()

GMeans

G-means

new GMeans()

Instance Members

▸ centroids

Centroids

centroids

Type: Array<Array<number>>

▸ size

Number of clusters.

size

Type: number

▸ clear()

Clear all clusters.

clear()

▸ fit(datas, iterations)

Fit model.

fit(datas: Array<Array<number>>, iterations: number)

Parameters

datas (Array<Array<number>>) Training data

iterations (number = -1) Iteration count

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

GMM

Gaussian mixture model

new GMM()

Instance Members

▸ add()

Add a new cluster.

add()

▸ clear()

Clear all clusters.

clear()

▸ probability(data)

Returns probabilities.

probability(data: Array<Array<number>>): Array<Array<number>>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

▸ predict(data)

Returns predicted categories.

predict(data: Array<Array<number>>): Array<number>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

▸ fit(datas)

Fit model.

fit(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

SemiSupervisedGMM

Semi-Supervised gaussian mixture model

new SemiSupervisedGMM()

Extends GMM

Instance Members

▸ categories

categories

Type: Array<any>

▸ init(datas, labels)

Initialize model.

init(datas: Array<Array<number>>, labels: Array<(any | null)>)

Parameters

datas (Array<Array<number>>) Training data

labels (Array<(any | null)>) Target values

▸ fit(datas, y)

Fit model.

fit(datas: Array<Array<number>>, y: Array<(any | null)>)

Parameters

datas (Array<Array<number>>) Training data

y (Array<(any | null)>) Target values

▸ predict(data)

Returns predicted categories.

predict(data: Array<Array<number>>): Array<any>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

GMR

Gaussian mixture regression

new GMR()

Extends GMM

Instance Members

▸ add()

Add a new cluster.

add()

▸ clear()

Clear all clusters.

clear()

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ probability(x, y)

Returns probabilities.

probability(x: Array<Array<number>>, y: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

y (Array<Array<number>>) Target values

Returns

Array<Array<number>>: Predicted values

▸ predict(x)

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

GPLVM

Gaussian Process Latent Variable Model

new GPLVM(rd: number, alpha: number, ez: number?, ea: number?, ep: number?, kernel: any)

Parameters

rd (number) Reduced dimension

alpha (number) Precision parameter

ez (number? = 1) Learning rate for z

ea (number? = 0.005) Learning rate for alpha

ep (number? = 0.2) Learning rate for kernel

kernel (any = 'gaussian')

Instance Members

▸ init(x)

Initialize model.

init(x: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

▸ fit()

Fit model.

fit()

▸ llh()

Returns log likelihood.

llh(): number

Returns

number: Log likelihood

▸ predict()

Returns reduced datas.

predict(): Array<Array<number>>

Returns

Array<Array<number>>: Predicted values

▸ reconstruct(z)

Returns reconstruct datas.

reconstruct(z: Array<Array<number>>): Array<Array<number>>

Parameters

z (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

GrowingCellStructures

Growing cell structures

new GrowingCellStructures()

Instance Members

▸ size

Number of clusters

size

Type: number

▸ update(x)

Update parameter.

update(x: Array<number>)

Parameters

x (Array<number>) Training data

▸ fit(x)

Fit model.

fit(x: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

GrowingNeuralGas

Growing neural gas

new GrowingNeuralGas(l: number, m: number)

Parameters

l (number) Neughborhood range

m (number) Decreasing factor of l

Instance Members

▸ size

Number of clusters

size

Type: number

▸ update(x)

Update parameter.

update(x: Array<number>)

Parameters

x (Array<number>) Training data

▸ fit(x)

Fit model.

fit(x: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

GSOM

Growing Self-Organizing Map

new GSOM(sf: number?, lr: number?)

Parameters

sf (number? = 0.1) Spread factor

lr (number? = 0.1) Learning rate

Instance Members

▸ size

Number of clusters

size

Type: number

▸ fit(data)

Fit model.

fit(data: Array<Array<number>>)

Parameters

data (Array<Array<number>>) Training data

▸ predict(x)

Returns predicted categories.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

GTM

Generative topographic mapping

new GTM(input_size: number, output_size: number, k: number?, q: number?)

Parameters

input_size (number) Input size

output_size (number) Output size

k (number? = 20) Grid size

q (number? = 10) Grid size for basis function

Instance Members

▸ probability(x)

Returns probabilities.

probability(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

▸ responsibility(x)

Returns responsibility.

responsibility(x: Array<Array<number>>): Matrix

Parameters

x (Array<Array<number>>) Sample data

Returns

Matrix: Responsibility

▸ fit(data)

Fit model.

fit(data: Array<Array<number>>)

Parameters

data (Array<Array<number>>) Training data

▸ predictIndex(x)

Returns best indexes.

predictIndex(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

▸ predict(x)

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

HamelryKMeans

Hamelry's accelerated k-Means algorithm

new HamelryKMeans(k: number)

Parameters

k (number) Number of clusters

Instance Members

▸ centroids

Centroids

centroids

Type: Array<Array<number>>

▸ init(datas)

Initialize this model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ fit()

Fit model.

fit()

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

HampelFilter

lib/model/hampel.js

Hampel filter

new HampelFilter(k: number?, th: number?)

Parameters

k (number? = 3) Half window size

th (number? = 3) Threshold

Instance Members

▸ predict(x)

lib/model/hampel.js

Returns predicted values.

predict(x: Array<number>): Array<number>

Parameters

x (Array<number>) Training data

Returns

Array<number>: Predicted values

HartiganWongKMeans

Hartigan-Wong k-Means algorithm

new HartiganWongKMeans(k: number)

Parameters

k (number) Number of clusters

Instance Members

▸ centroids

Centroids

centroids

Type: Array<Array<number>>

▸ init(datas)

Initialize this model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ fit()

Fit model.

fit()

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

HDBSCAN

lib/model/hdbscan.js

Hierarchical Density-based spatial clustering of applications with noise

new HDBSCAN(minClusterSize: number?, minPts: number?, metric: ("euclid" | "manhattan" | "chebyshev" | function (Array<number>, Array<number>): number)?)

Parameters

minClusterSize (number? = 5) Minimum number of clusters to be recognized as a cluster

minPts (number? = 5) Number of neighborhood with core distance

metric (("euclid" | "manhattan" | "chebyshev" | function (Array<number>, Array<number>): number)? = 'euclid') Metric name

Instance Members

▸ size

lib/model/hdbscan.js

Number of clusters of last predicted

size

Type: number

▸ predict(datas)

lib/model/hdbscan.js

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Predicted values

Histogram

lib/model/histogram.js

Histogram

new Histogram(config: object?)

Parameters

config (object? = {}) Config

Instance Members

▸ fit(datas)

lib/model/histogram.js

Returns histogram data.

fit(datas: Array<Array<number>>): Array<any>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<any>: Predicted values. An array nested by the number of dimensions of the data

▸ predict(datas)

lib/model/histogram.js

Returns predicted counted values.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

HLLE

lib/model/hlle.js

Hessian Locally Linear Embedding

new HLLE(k: number, rd: (number | null)?)

Parameters

k (number) Number of neighborhoods

rd ((number | null)? = null) Reduced dimension

Instance Members

▸ predict(x)

lib/model/hlle.js

Returns reduced values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<number>>: Predicted values

HMMBase

Hidden Markov model

new HMMBase(n: number)

Parameters

n (number) Number of states

Instance Members

▸ probability(x)

Returns probability of the datas.

probability(x: Matrix): Matrix

Parameters

x (Matrix) Sample data

Returns

Matrix: Predicted values

▸ bestPath(x)

Returns best path of the datas.

bestPath(x: Matrix): Matrix

Parameters

x (Matrix) Sample data

Returns

Matrix: Predicted path

HMM

Hidden Markov model

new HMM(n: number)

Extends HMMBase

Parameters

n (number) Number of states

Instance Members

▸ fit(datas, scaled)

Fit model.

fit(datas: Array<Array<any>>, scaled: boolean)

Parameters

datas (Array<Array<any>>) Training data

scaled (boolean = false) Do scaled calculation or not

▸ probability(datas)

Returns probability of the datas.

probability(datas: Array<Array<any>>): Array<number>

Parameters

datas (Array<Array<any>>) Sample data

Returns

Array<number>: Predicted values

▸ bestPath(data)

Returns best path of the datas.

bestPath(data: Array<Array<any>>): Array<Array<number>>

Parameters

data (Array<Array<any>>) Sample data

Returns

Array<Array<number>>: Predicted path

ContinuousHMM

Continuous hidden Markov model

new ContinuousHMM(n: number)

Extends HMMBase

Parameters

n (number) Number of states

Instance Members

▸ fit(x, scaled)

Fit model.

fit(x: (Array<Array<number>> | Array<Array<Array<number>>>), scaled: boolean)

Parameters

x ((Array<Array<number>> | Array<Array<Array<number>>>)) Training data

scaled (boolean = false) Do scaled calculation or not

▸ probability(datas)

Returns probability of the datas.

probability(datas: (Array<Array<number>> | Array<Array<Array<number>>>)): Array<number>

Parameters

datas ((Array<Array<number>> | Array<Array<Array<number>>>)) Sample data

Returns

Array<number>: Predicted values

▸ bestPath(data)

Returns best path of the datas.

bestPath(data: (Array<Array<number>> | Array<Array<Array<number>>>)): Array<Array<number>>

Parameters

data ((Array<Array<number>> | Array<Array<Array<number>>>)) Sample data

Returns

Array<Array<number>>: Predicted path

▸ generate(n, length)

lib/model/holt_winters.js

Returns generated values.

generate(n: number, length: number): Array<Array<Array<number>>>

Parameters

n (number = 1) Number of generated data

length (number = 5) Path length

Returns

Array<Array<Array<number>>>: Generated values

HoltWinters

Holt-Winters method

new HoltWinters(a: number, b: number?, g: number?, s: number?)

Parameters

a (number) Weight for last value

b (number? = 0) Weight for trend value

g (number? = 0) Weight for seasonal data

s (number? = 0) Length of season

Instance Members

▸ fit(x)

lib/model/holt_winters.js

Fit model and return predict values.

fit(x: Array<number>): Array<number>

Parameters

x (Array<number>) Training data

Returns

Array<number>: Predicted values

▸ predict(k)

lib/model/holt_winters.js

Returns predicted future values.

predict(k: number): Array<Array<number>>

Parameters

k (number) Prediction count

Returns

Array<Array<number>>: Predicted values

HopfieldNetwork

Hopfield network

new HopfieldNetwork()

Instance Members

▸ fit(x)

Fit model.

fit(x: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

▸ energy(x)

Return a energy value of the data.

energy(x: Array<number>): number

Parameters

x (Array<number>) Sample data

Returns

number: Energy value

▸ predict(x)

lib/model/huber_regression.js

Returns predicted values.

predict(x: Array<number>): Array<number>

Parameters

x (Array<number>) Sample data

Returns

Array<number>: Predicted values

Hotelling

lib/model/hotelling.js

Hotelling T-square Method

new Hotelling()

Instance Members

▸ fit(data)

lib/model/hotelling.js

Fit model.

fit(data: Array<Array<number>>)

Parameters

data (Array<Array<number>>) Training data

▸ predict(data)

lib/model/hotelling.js

Returns anomaly degrees.

predict(data: Array<Array<number>>): Array<number>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

HuberRegression

Huber regression

new HuberRegression(e: number?, method: ("rls" | "gd")?, lr: number?)

Parameters

e (number? = 1.35) Threshold of outliers

method (("rls" | "gd")? = 'rls') Method name

lr (number? = 1) Learning rate

Instance Members

▸ fit(x, y)

lib/model/huber_regression.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

lib/model/huber_regression.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

ICA

lib/model/ica.js

Independent component analysis

new ICA(rd: (number | null)?)

Parameters

rd ((number | null)? = null) Reduced dimension

Instance Members

▸ fit(x)

lib/model/ica.js

Fit model.

fit(x: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

▸ predict(x)

lib/model/ica.js

Returns reduced datas.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<number>>: Predicted values

CELLIP

Classical ellipsoid method

new CELLIP(gamma: number?, a: number?)

Parameters

gamma (number? = 0.1) Desired classification margin

a (number? = 0.5) Tradeoff parameter

Instance Members

▸ init(train_x, train_y)

Initialize this model.

init(train_x: Array<Array<number>>, train_y: Array<(1 | -1)>)

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<(1 | -1)>) Target values

▸ update(x, y)

Update model parameters with one data.

update(x: Matrix, y: (1 | -1))

Parameters

x (Matrix) Training data

y ((1 | -1)) Target value

▸ fit()

Fit model parameters.

fit()

▸ predict(data)

Returns predicted datas.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

IELLIP

Improved ellipsoid method

new IELLIP(b: number?, c: number?)

Parameters

b (number? = 0.9) Parameter controlling the memory of online learning

c (number? = 0.5) Parameter controlling the memory of online learning

Instance Members

▸ init(train_x, train_y)

Initialize this model.

init(train_x: Array<Array<number>>, train_y: Array<(1 | -1)>)

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<(1 | -1)>) Target values

▸ update(x, y)

Update model parameters with one data.

update(x: Matrix, y: (1 | -1))

Parameters

x (Matrix) Training data

y ((1 | -1)) Target value

▸ fit()

Fit model parameters.

fit()

▸ predict(data)

Returns predicted datas.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

IKNN

lib/model/iknn.js

Locally Informative K-Nearest Neighbor

new IKNN(k: number, i: number)

Parameters

k (number) Number of neighbors

i (number) Number of informative points

Instance Members

▸ fit(x, y)

lib/model/iknn.js

Fit model.

fit(x: Array<Array<number>>, y: Array<any>)

Parameters

x (Array<Array<number>>) Training data

y (Array<any>) Target values

▸ predict(data)

lib/model/iknn.js

Returns predicted categories.

predict(data: Array<Array<number>>): Array<any>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

IncrementalPCA

Incremental principal component analysis

new IncrementalPCA(f: number?, rd: (number | null)?)

Parameters

f (number? = 0.95) Forgetting factor

rd ((number | null)? = null) Reduced dimension

Instance Members

▸ update(x)

Update parameters.

update(x: Matrix)

Parameters

x (Matrix) Training data

▸ fit(x)

Fit model.

fit(x: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

▸ predict(x)

lib/model/inverse_distance_weighting.js

Returns reduced datas.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

INFLO

lib/model/inflo.js

Influenced Outlierness

new INFLO(k: number)

Parameters

k (number) Number of neighborhoods

Instance Members

▸ predict(datas)

lib/model/inflo.js

Returns anomaly degrees.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Predicted values

InverseDistanceWeighting

Inverse distance weighting

new InverseDistanceWeighting(k: number?, p: number?, metric: ("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)?)

Parameters

k (number? = 5) Number of neighborhoods

p (number? = 2) Power parameter

metric (("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)? = 'euclid') Metric name

Instance Members

▸ fit(x, y)

lib/model/inverse_distance_weighting.js

Fit model.

fit(x: Array<Array<number>>, y: Array<number>)

Parameters

x (Array<Array<number>>) Training data

y (Array<number>) Target values

▸ predict(data)

lib/model/inverse_distance_weighting.js

Returns predicted values.

predict(data: Array<Array<number>>): Array<number>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

InverseSmoothstepInterpolation

lib/model/inverse_smoothstep.js

Inverse smoothstep interpolation

new InverseSmoothstepInterpolation()

Instance Members

▸ fit(x, y)

lib/model/inverse_smoothstep.js

Fit model.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(target)

lib/model/inverse_smoothstep.js

Returns predicted interpolated values.

predict(target: Array<number>): Array<number>

Parameters

target (Array<number>) Sample data

Returns

Array<number>: Predicted values

ISODATA

Iterative Self-Organizing Data Analysis Technique

new ISODATA(init_k: number, min_k: number, max_k: number, min_n: number, split_std: number, merge_dist: number)

Parameters

init_k (number) Initial cluster count

min_k (number) Minimum cluster count

max_k (number) Maximum cluster count

min_n (number) Minimum cluster size

split_std (number) Standard deviation as splid threshold

merge_dist (number) Merge distance

Instance Members

▸ centroids

Centroids

centroids

Type: Array<Array<number>>

▸ size

Number of clusters

size

Type: number

▸ init(data)

Initialize model.

init(data: Array<Array<number>>)

Parameters

data (Array<Array<number>>) Training data

▸ fit(data)

Fit model.

fit(data: Array<Array<number>>)

Parameters

data (Array<Array<number>>) Training data

▸ predict(datas)

lib/model/isolation_forest.js

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

IsolationForest

Isolation forest

new IsolationForest(tree_num: number?, sampling_rate: number?)

Parameters

tree_num (number? = 100) Number of trees

sampling_rate (number? = 0.8) Sampling rate

Instance Members

▸ fit(datas)

lib/model/isolation_forest.js

Fit model.

fit(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ predict(datas)

lib/model/isolation_forest.js

Returns anomaly degrees.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

Isomap

lib/model/isomap.js

Isomap

new Isomap(neighbors: number?, rd: (number | null)?)

Parameters

neighbors (number? = 0) Number of neighborhoods

rd ((number | null)? = null) Reduced dimension

Instance Members

▸ predict(x)

lib/model/isomap.js

Returns reduced values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<number>>: Predicted values

IsotonicRegression

lib/model/isotonic.js

Isotonic regression

new IsotonicRegression()

Instance Members

▸ fit(x, y)

lib/model/isotonic.js

Fit model.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(x)

lib/model/isotonic.js

Returns predicted values.

predict(x: Array<number>): Array<number>

Parameters

x (Array<number>) Sample data

Returns

Array<number>: Predicted values

KalmanFilter

lib/model/kalman_filter.js

Kalman filter

new KalmanFilter()

Instance Members

▸ fit(z)

lib/model/kalman_filter.js

Fit and returns smoothed values.

fit(z: Array<Array<number>>): Array<Array<number>>

Parameters

z (Array<Array<number>>) Training data

Returns

Array<Array<number>>: Predicted values

▸ predict(k)

lib/model/kalman_filter.js

Returns predicted future values.

predict(k: number): Array<Array<number>>

Parameters

k (number) Prediction count

Returns

Array<Array<number>>: Predicted values

KDEOS

lib/model/kdeos.js

Kernel Density Estimation Outlier Score

new KDEOS(kmin: number, kmax: number, kernel: ("gaussian" | "epanechnikov" | {name: "gaussian"} | {name: "epanechnikov"} | function (number, number, number): number)?)

Parameters

kmin (number) Minimum number of neighborhoods

kmax (number) Maximum number of neighborhoods

kernel

(("gaussian" | "epanechnikov" | {name: "gaussian"} | {name: "epanechnikov"} | function (number, number, number): number)? = 'gaussian')

Kernel name

Instance Members

▸ predict(datas)

lib/model/kdeos.js

Returns anomaly degrees.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Predicted values

KernelDensityEstimator

Kernel density estimator

new KernelDensityEstimator(h: number?, kernel: ("gaussian" | "rectangular" | "triangular" | "epanechnikov" | "biweight" | "triweight" | {name: "gaussian"} | {name: "rectangular"} | {name: "triangular"} | {name: "epanechnikov"} | {name: "biweight"} | {name: "triweight"} | function (number): number)?)

Parameters

h (number? = 0) Smoothing parameter for the kernel

kernel

(("gaussian" | "rectangular" | "triangular" | "epanechnikov" | "biweight" | "triweight" | {name: "gaussian"} | {name: "rectangular"} | {name: "triangular"} | {name: "epanechnikov"} | {name: "biweight"} | {name: "triweight"} | function (number): number)? = 'gaussian')

Kernel name

Instance Members

▸ fit(x)

Fit model.

fit(x: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

▸ probability(x)

Returns probabilities of the datas.

probability(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

▸ predict(x)

Returns probabilities of the datas.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

KernelKMeans

Kernel k-means

new KernelKMeans(k: number?)

Parameters

k (number? = 3) Number of clusters

Instance Members

▸ init(datas)

Initialize model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ predict()

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

▸ fit()

Fit model.

fit()

KernelizedPegasos

Kernelized Primal Estimated sub-GrAdientSOlver for SVM

new KernelizedPegasos(rate: number, kernel: any)

Parameters

rate (number) Learning rate

kernel (any = 'gaussian')

Instance Members

▸ init(train_x, train_y)

Initialize this model.

init(train_x: Array<Array<number>>, train_y: Array<(1 | -1)>)

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<(1 | -1)>) Target values

▸ fit()

Fit model parameters.

fit()

▸ predict(data)

lib/model/kernelized_perceptron.js

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

KernelizedPerceptron

Kernelized perceptron

new KernelizedPerceptron(rate: number?, kernel: any)

Parameters

rate (number? = 1) Learning rate

kernel (any = 'gaussian')

Instance Members

▸ fit(x, y)

lib/model/kernelized_perceptron.js

Fit model.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

lib/model/kernelized_perceptron.js

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

KHarmonicMeans

K-Harmonic Means

new KHarmonicMeans(k: number)

Parameters

k (number) Number of clusters

Instance Members

▸ centroids

Centroids

centroids

Type: Array<Array<number>>

▸ init(datas)

Initialize this model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ fit()

Fit model.

fit()

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

KLIEP

lib/model/kliep.js

Kullback-Leibler importance estimation procedure

new KLIEP(sigma: Array<number>, fold: number, kernelNum: number)

Parameters

sigma (Array<number>) Sigmas of normal distribution

fold (number) Number of folds

kernelNum (number) Number of kernels

Instance Members

▸ fit(x1, x2)

lib/model/kliep.js

Fit model.

fit(x1: Array<Array<number>>, x2: Array<Array<number>>)

Parameters

x1 (Array<Array<number>>) Numerator data

x2 (Array<Array<number>>) Denominator data

▸ predict(x)

lib/model/kliep.js

Returns estimated values.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

KMeansBase

Bsae class for k-means like model

new KMeansBase()

Instance Members

▸ centroids

Centroids

centroids

Type: Array<Array<number>>

▸ size

Number of clusters.

size

Type: number

▸ add(datas)

Add a new cluster.

add(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Added centroid

▸ clear()

Clear all clusters.

clear()

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

▸ fit(datas)

Fit model and returns total distance the centroid has moved.

fit(datas: Array<Array<number>>): number

Parameters

datas (Array<Array<number>>) Training data

Returns

number: Total distance the centroid has moved

KMeans

k-means model

new KMeans()

Extends KMeansBase

Instance Members

▸ _add(centroids, datas)

Returns a new centroid.

_add(centroids: Array<Array<number>>, datas: Array<Array<number>>): Array<number>

Parameters

centroids (Array<Array<number>>) Centroids

datas (Array<Array<number>>) Training data

Returns

Array<number>: Added centroid

▸ _move(centroids, datas)

Returns moved centroid positions.

_move(centroids: Array<Array<number>>, datas: Array<Array<number>>): Array<Array<number>>

Parameters

centroids (Array<Array<number>>) Centroids

datas (Array<Array<number>>) Training data

Returns

Array<Array<number>>: Moved centroids

KMeanspp

k-means++ model

new KMeanspp()

Extends KMeans

Instance Members

▸ _add(centroids, datas)

Returns a new centroid.

_add(centroids: Array<Array<number>>, datas: Array<Array<number>>): Array<number>

Parameters

centroids (Array<Array<number>>) Centroids

datas (Array<Array<number>>) Training data

Returns

Array<number>: Added centroid

KMedoids

k-medoids model

new KMedoids()

Extends KMeans

Instance Members

▸ _move(centroids, datas)

Returns moved centroid positions.

_move(centroids: Array<Array<number>>, datas: Array<Array<number>>): Array<Array<number>>

Parameters

centroids (Array<Array<number>>) Centroids

datas (Array<Array<number>>) Training data

Returns

Array<Array<number>>: Moved centroids

KMedians

k-medians model

new KMedians()

Extends KMeans

Instance Members

▸ _move(centroids, datas)

Returns moved centroid positions.

_move(centroids: Array<Array<number>>, datas: Array<Array<number>>): Array<Array<number>>

Parameters

centroids (Array<Array<number>>) Centroids

datas (Array<Array<number>>) Training data

Returns

Array<Array<number>>: Moved centroids

SemiSupervisedKMeansModel

semi-supervised k-means model

new SemiSupervisedKMeansModel()

Extends KMeansBase

Instance Members

▸ categories

categories

Type: Array<any>

▸ init(datas, labels)

Initialize model.

init(datas: Array<Array<number>>, labels: Array<(any | null)>)

Parameters

datas (Array<Array<number>>) Training data

labels (Array<(any | null)>) Target values

▸ fit(datas, labels)

Fit and returns total distance the centroid has moved.

fit(datas: Array<Array<number>>, labels: Array<(any | null)>): number

Parameters

datas (Array<Array<number>>) Training data

labels (Array<(any | null)>) Target values

Returns

number: Total distance the centroid has moved

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<any>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

KModes

k-modes model

new KModes()

Instance Members

▸ size

Number of clusters.

size

Type: number

▸ add(datas)

Add a new cluster.

add(datas: Array<Array<any>>)

Parameters

datas (Array<Array<any>>) Training data

▸ clear()

Clear all clusters.

clear()

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<any>>): Array<number>

Parameters

datas (Array<Array<any>>) Sample data

Returns

Array<number>: Predicted values

▸ fit(datas)

Fit model and returns total distance the modes has moved.

fit(datas: Array<Array<any>>): number

Parameters

datas (Array<Array<any>>) Training data

Returns

number: Total distance the modes has moved

KNNBase

Bsae class for k-nearest neighbor models

new KNNBase(k: number?, metric: ("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)?)

Parameters

k (number? = 5) Number of neighborhoods

metric (("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)? = 'euclid') Metric name

Instance Members

▸ _add(point, category?)

Add a data.

_add(point: Array<number>, category: any?)

Parameters

point (Array<number>) Training data

category (any?) Target value

KNN

k-nearest neighbor

new KNN(k: number?, metric: ("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)?)

Extends KNNBase

Parameters

k (number? = 5) Number of neighborhoods

metric (("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)? = 'euclid') Metric name

Instance Members

▸ add(point, category)

Add a data.

add(point: Array<number>, category: any)

Parameters

point (Array<number>) Training data

category (any) Target value

▸ fit(datas, targets)

Add datas.

fit(datas: Array<Array<number>>, targets: Array<any>)

Parameters

datas (Array<Array<number>>) Training data

targets (Array<any>) Target values

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<any>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

KNNRegression

k-nearest neighbor regression

new KNNRegression(k: number?, metric: ("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)?)

Extends KNNBase

Parameters

k (number? = 5) Number of neighborhoods

metric (("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)? = 'euclid') Metric name

Instance Members

▸ add(point, category)

Add a data.

add(point: Array<number>, category: number)

Parameters

point (Array<number>) Training data

category (number) Target value

▸ fit(datas, targets)

Add datas.

fit(datas: Array<Array<number>>, targets: Array<number>)

Parameters

datas (Array<Array<number>>) Training data

targets (Array<number>) Target values

▸ predict(datas)

Returns predicted values.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

KNNAnomaly

k-nearest neighbor anomaly detection

new KNNAnomaly(k: number?, metric: ("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)?)

Extends KNNBase

Parameters

k (number? = 5) Number of neighborhoods

metric (("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)? = 'euclid') Metric name

Instance Members

▸ add(point)

Add a data.

add(point: Array<number>)

Parameters

point (Array<number>) Training data

▸ fit(datas)

Add datas.

fit(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ predict(datas)

Returns anomaly degrees.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

KNNDensityEstimation

k-nearest neighbor density estimation

new KNNDensityEstimation(k: number?, metric: ("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)?)

Extends KNNBase

Parameters

k (number? = 5) Number of neighborhoods

metric (("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)? = 'euclid') Metric name

Instance Members

▸ add(point)

Add a data.

add(point: Array<number>)

Parameters

point (Array<number>) Training data

▸ fit(datas)

Add datas.

fit(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ predict(datas)

Returns predicted values.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

SemiSupervisedKNN

Semi-supervised k-nearest neighbor

new SemiSupervisedKNN(k: number?, metric: ("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)?)

Extends KNNBase

Parameters

k (number? = 5) Number of neighborhoods

metric (("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)? = 'euclid') Metric name

Instance Members

▸ add(point, category)

Add a data.

add(point: Array<number>, category: (any | null))

Parameters

point (Array<number>) Training data

category ((any | null)) Target value

▸ fit(datas, targets)

Add datas.

fit(datas: Array<Array<number>>, targets: Array<(any | null)>)

Parameters

datas (Array<Array<number>>) Training data

targets (Array<(any | null)>) Target values

▸ predict()

Returns predicted values.

predict(): Array<any>

Returns

Array<any>: Predicted values

KPrototypes

k-prototypes model

new KPrototypes(gamma: number, categoryPositions: Array<boolean>)

Parameters

gamma (number) Weight for categorical data

categoryPositions (Array<boolean>) Category column position

Instance Members

▸ size

Number of clusters.

size

Type: number

▸ add(datas)

Add a new cluster.

add(datas: Array<Array<any>>)

Parameters

datas (Array<Array<any>>) Training data

▸ clear()

Clear all clusters.

clear()

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<any>>): Array<number>

Parameters

datas (Array<Array<any>>) Sample data

Returns

Array<number>: Predicted values

▸ fit(datas)

Fit model and returns total distance the modes has moved.

fit(datas: Array<Array<any>>): number

Parameters

datas (Array<Array<any>>) Training data

Returns

number: Total distance the modes has moved

KSVD

lib/model/ksvd.js

k-SVD

new KSVD(x: Array<Array<number>>, m: number, k: number?)

Parameters

x (Array<Array<number>>) Training data

m (number) Reduced dimension

k (number? = m) Sparsity parameter

Instance Members

▸ fit()

lib/model/ksvd.js

Fit model and returns error.

fit(): number

Returns

number: Error

▸ predict()

lib/model/ksvd.js

Returns reduced values.

predict(): Array<Array<number>>

Returns

Array<Array<number>>: Predicted values

KolmogorovZurbenkoFilter

lib/model/kz.js

Kolmogorov–Zurbenko filter

new KolmogorovZurbenkoFilter(m: number, k: number)

Parameters

m (number) Window size

k (number) Iteration count of a moving average

Instance Members

▸ predict(x)

lib/model/kz.js

Returns predicted values.

predict(x: Array<number>): Array<number>

Parameters

x (Array<number>) Training data

Returns

Array<number>: Predicted values

LabelPropagation

Label propagation

new LabelPropagation(method: ("rbf" | "knn")?, sigma: number?, k: number?)

Parameters

method (("rbf" | "knn")? = 'rbf') Method name

sigma (number? = 0.1) Sigma of normal distribution

k (number? = Infinity) Number of neighborhoods

Instance Members

▸ init(x, y)

Initialize model.

init(x: Array<Array<number>>, y: Array<(any | null)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(any | null)>) Target values

▸ fit()

Fit model.

fit()

▸ predict()

Returns predicted categories.

predict(): Array<any>

Returns

Array<any>: Predicted values

LabelSpreading

Label spreading

new LabelSpreading(alpha: number?, method: ("rbf" | "knn")?, sigma: number?, k: number?)

Parameters

alpha (number? = 0.2) Clamping factor

method (("rbf" | "knn")? = 'rbf') Method name

sigma (number? = 0.1) Sigma of normal distribution

k (number? = Infinity) Number of neighborhoods

Instance Members

▸ init(x, y)

Initialize model.

init(x: Array<Array<number>>, y: Array<(any | null)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(any | null)>) Target values

▸ fit()

Fit model.

fit()

▸ predict()

Returns predicted categories.

predict(): Array<any>

Returns

Array<any>: Predicted values

LadderNetwork

Ladder network

new LadderNetwork(hidden_sizes: Array<number>, lambdas: Array<number>, activation: string, optimizer: string)

Parameters

hidden_sizes (Array<number>) Sizes of hidden layers

lambdas (Array<number>) Regularization parameters

activation (string) Activation name

optimizer (string) Optimizer of the network

Instance Members

▸ epoch

Epoch

epoch

Type: number

▸ fit(train_x, train_y, iteration, rate, batch)

Fit model.

fit(train_x: Array<Array<number>>, train_y: Array<(any | null)>, iteration: number, rate: number, batch: number): {labeledLoss: number, unlabeledLoss: number}

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<(any | null)>) Target values

iteration (number) Iteration count

rate (number) Learning rate

batch (number) Batch size

Returns

{labeledLoss: number, unlabeledLoss: number}: Loss value

▸ predict(x)

lib/model/lanczos_interpolation.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<any>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

LagrangeInterpolation

lib/model/lagrange.js

Lagrange interpolation

new LagrangeInterpolation(method: ("weighted" | "newton" | "")?)

Parameters

method (("weighted" | "newton" | "")? = 'weighted') Method name

Instance Members

▸ fit(x, y)

lib/model/lagrange.js

Fit model.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(target)

lib/model/lagrange.js

Returns predicted interpolated values.

predict(target: Array<number>): Array<number>

Parameters

target (Array<number>) Sample data

Returns

Array<number>: Predicted values

LanczosInterpolation

Lanczos interpolation

new LanczosInterpolation(n: number)

Parameters

n (number) Order

Instance Members

▸ fit(values)

lib/model/lanczos_interpolation.js

Fit model parameters.

fit(values: Array<number>)

Parameters

values (Array<number>) Training data

▸ predict(index)

lib/model/lanczos_interpolation.js

Returns predicted interpolated values.

predict(index: Array<number>): Array<number>

Parameters

index (Array<number>) Training data

Returns

Array<number>: Predicted values

Laplacian

lib/model/laplacian.js

Laplacian edge detection

new Laplacian(th: number, n: (4 | 8)?)

Parameters

th (number) Threshold

n ((4 | 8)? = 4) Number of neighborhoods

Instance Members

▸ predict(x)

lib/model/laplacian.js

Returns predicted edge flags.

predict(x: Array<Array<number>>): Array<Array<boolean>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<boolean>>: Predicted values. true if a pixel is edge.

LaplacianEigenmaps

lib/model/laplacian_eigenmaps.js

Laplacian eigenmaps

new LaplacianEigenmaps(rd: number, affinity: ("rbf" | "knn")?, k: number?, sigma: number?, laplacian: ("unnormalized" | "normalized")?)

Parameters

rd (number) Reduced dimension

affinity (("rbf" | "knn")? = 'rbf') Affinity type name

k (number? = 10) Number of neighborhoods

sigma (number? = 1) Sigma of normal distribution

laplacian (("unnormalized" | "normalized")? = 'unnormalized') Normalized laplacian matrix or not

Instance Members

▸ predict(x)

lib/model/laplacian_eigenmaps.js

Returns reduced datas.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<number>>: Predicted values

Lasso

Least absolute shrinkage and selection operator

new Lasso(lambda: number?, method: ("CD" | "ISTA" | "LARS")?)

Parameters

lambda (number? = 0.1) Regularization strength

method (("CD" | "ISTA" | "LARS")? = 'CD') Method name

Instance Members

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

▸ importance()

Returns importances of the features.

importance(): Array<number>

Returns

Array<number>: Importances

LatentDirichletAllocation

Latent dirichlet allocation

new LatentDirichletAllocation(t: number?)

Parameters

t (number? = 2) Topic count

Instance Members

▸ init(x)

Initialize model.

init(x: Array<Array<any>>)

Parameters

x (Array<Array<any>>) Training data

▸ fit()

Fit model.

fit()

▸ predict()

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

LBG

Linde-Buzo-Gray algorithm

new LBG()

Instance Members

▸ centroids

Centroids

centroids

Type: Array<Array<number>>

▸ size

Number of clusters.

size

Type: number

▸ clear()

Clear all clusters.

clear()

▸ fit(datas)

Fit model.

fit(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

LinearDiscriminant

Linear discriminant analysis

new LinearDiscriminant()

Instance Members

▸ init(train_x, train_y)

Initialize this model.

init(train_x: Array<Array<number>>, train_y: Array<(1 | -1)>)

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<(1 | -1)>) Target values

▸ fit()

Fit model parameters.

fit()

▸ predict(data)

Returns predicted datas.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

FishersLinearDiscriminant

Fishers linear discriminant analysis

new FishersLinearDiscriminant()

Instance Members

▸ init(train_x, train_y)

Initialize this model.

init(train_x: Array<Array<number>>, train_y: Array<(1 | -1)>)

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<(1 | -1)>) Target values

▸ fit()

Fit model parameters.

fit()

▸ predict(data)

Returns predicted datas.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

MulticlassLinearDiscriminant

Multiclass linear discriminant analysis

new MulticlassLinearDiscriminant()

Instance Members

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<any>)

Parameters

x (Array<Array<number>>) Training data

y (Array<any>) Target values

▸ predict(data)

Returns predicted categories.

predict(data: Array<Array<number>>): Array<any>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

LinearDiscriminantAnalysis

Linear discriminant analysis

new LinearDiscriminantAnalysis(rd: (number | null)?)

Parameters

rd ((number | null)? = null) Reduced dimension

Instance Members

▸ predict(x, t)

lib/model/least_absolute.js

Returns reduced values.

predict(x: Array<Array<number>>, t: Array<number>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Training data

t (Array<number>) Target values

Returns

Array<Array<number>>: Predicted values

LDF

lib/model/ldf.js

Local Density Factor

new LDF(k: number)

Parameters

k (number) Number of neighborhoods

Instance Members

▸ predict(datas)

lib/model/ldf.js

Returns anomaly degrees.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Predicted values

LDOF

lib/model/ldof.js

Local Distance-based Outlier Factor

new LDOF(k: number)

Parameters

k (number) Number of neighborhoods

Instance Members

▸ predict(datas)

lib/model/ldof.js

Returns anomaly degrees.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Predicted values

LeastAbsolute

Least absolute deviations

new LeastAbsolute()

Instance Members

▸ fit(x, y)

lib/model/least_absolute.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

lib/model/least_absolute.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

LeastSquares

lib/model/least_square.js

Least squares

new LeastSquares()

Instance Members

▸ fit(x, y)

lib/model/least_square.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

lib/model/least_square.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

LinearInterpolation

lib/model/lerp.js

Linear interpolation

new LinearInterpolation()

Instance Members

▸ fit(x, y)

lib/model/lerp.js

Fit model parameters.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(target)

lib/model/lerp.js

Returns predicted interpolated values.

predict(target: Array<number>): Array<number>

Parameters

target (Array<number>) Sample data

Returns

Array<number>: Predicted values

LLE

lib/model/lle.js

Locally Linear Embedding

new LLE(k: number?, rd: (number | null)?)

Parameters

k (number? = 1) Number of neighborhoods

rd ((number | null)? = null) Reduced dimension

Instance Members

▸ predict(x)

lib/model/lle.js

Returns reduced values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<number>>: Predicted values

LeastMedianSquaresRegression

lib/model/lmeds.js

Least median squares regression

new LeastMedianSquaresRegression(k: number)

Parameters

k (number = 5) Sampling count

Instance Members

▸ fit(x, y)

lib/model/lmeds.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

lib/model/lmeds.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

LMNN

Large Margin Nearest Neighbor

new LMNN(gamma: number, lambda: number)

Parameters

gamma (number) Tuning parameter

lambda (number) Tuning parameter

Instance Members

▸ init(x, y)

Initialize model.

init(x: Array<Array<number>>, y: Array<any>)

Parameters

x (Array<Array<number>>) Training data

y (Array<any>) Target values

▸ fit()

Fit model.

fit()

▸ predict(x)

lib/model/logarithmic_interpolation.js

Returns predicted categories.

predict(x: Array<Array<number>>): Array<any>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

LOCI

lib/model/loci.js

Local Correlation Integral

new LOCI(alpha: number?)

Parameters

alpha (number? = 0.5) Alpha

Instance Members

▸ predict(datas)

lib/model/loci.js

Returns a list of the data predicted as outliers or not.

predict(datas: Array<Array<number>>): Array<boolean>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<boolean>: Predicted values

LOESS

lib/model/loess.js

Locally estimated scatterplot smoothing

new LOESS()

Instance Members

▸ fit(x, y)

lib/model/loess.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

lib/model/loess.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

LOF

lib/model/lof.js

Local Outlier Factor

new LOF(k: number)

Parameters

k (number) Number of neighborhoods

Instance Members

▸ predict(datas)

lib/model/lof.js

Returns anomaly degrees.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Predicted values

LoG

lib/model/log.js

Laplacian of gaussian filter

new LoG(th: number)

Parameters

th (number) Threshold

Instance Members

▸ predict(x)

lib/model/log.js

Returns predicted edge flags.

predict(x: Array<Array<number>>): Array<Array<boolean>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<boolean>>: Predicted values. true if a pixel is edge.

LogarithmicInterpolation

Logarithmic interpolation

new LogarithmicInterpolation()

Instance Members

▸ fit(x, y)

lib/model/logarithmic_interpolation.js

Fit model.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(target)

lib/model/logarithmic_interpolation.js

Returns predicted interpolated values.

predict(target: Array<number>): Array<number>

Parameters

target (Array<number>) Sample data

Returns

Array<number>: Predicted values

LogisticRegression

Logistic regression

new LogisticRegression()

Instance Members

▸ fit(x, y, iteration = 1, rate = 0.1, l1 = 0, l2 = 0)

Fit model.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>, iteration: number?, rate: number?, l1: number?, l2: number?)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

iteration (number? = 1) Iteration count

rate (number? = 0.1) Learning rate

l1 (number? = 0) L1 regularization strength

l2 (number? = 0) L2 regularization strength

▸ predict(points)

Returns predicted values.

predict(points: Array<Array<number>>): Array<(1 | -1)>

Parameters

points (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

MultinomialLogisticRegression

Multinomial logistic regression

new MultinomialLogisticRegression(classes: Array<number>?)

Parameters

classes (Array<number>?) Initial class labels

Instance Members

▸ fit(train_x, train_y, iteration = 1, rate = 0.1, l1 = 0, l2 = 0)

Fit model.

fit(train_x: Array<Array<number>>, train_y: Array<any>, iteration: number?, rate: number?, l1: number?, l2: number?)

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<any>) Target values

iteration (number? = 1) Iteration count

rate (number? = 0.1) Learning rate

l1 (number? = 0) L1 regularization strength

l2 (number? = 0) L2 regularization strength

▸ predict(points)

lib/model/lpnorm_linear.js

Returns predicted categories.

predict(points: Array<Array<number>>): Array<any>

Parameters

points (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

LoOP

lib/model/loop.js

Local Outlier Probability

new LoOP(k: number)

Parameters

k (number) Number of neighborhoods

Instance Members

▸ predict(datas)

lib/model/loop.js

Returns anomaly degrees.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Predicted values

LOWESS

lib/model/lowess.js

Locally weighted scatter plot smooth

new LOWESS()

Instance Members

▸ fit(x, y)

lib/model/lowess.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

lib/model/lowess.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

LowpassFilter

lib/model/lowpass.js

Lowpass filter

new LowpassFilter(c: number?)

Parameters

c (number? = 0.5) Cutoff rate

Instance Members

▸ predict(x)

lib/model/lowpass.js

Returns predicted datas.

predict(x: Array<number>): Array<number>

Parameters

x (Array<number>) Training data

Returns

Array<number>: Predicted values

LpNormLinearRegression

Lp norm linear regression

new LpNormLinearRegression(p: number?)

Parameters

p (number? = 2) Power parameter for norm

Instance Members

▸ fit(x, y)

lib/model/lpnorm_linear.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

lib/model/lpnorm_linear.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

LSA

lib/model/lsa.js

Latent Semantic Analysis

new LSA(rd: (number | null)?)

Parameters

rd ((number | null)? = null) Reduced dimension

Instance Members

▸ predict(x)

lib/model/lsa.js

Returns reduced values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<number>>: Predicted values

LSDD

Least-squares density difference

new LSDD(sigma: Array<number>, lambda: Array<number>)

Parameters

sigma (Array<number>) Sigmas of normal distribution

lambda (Array<number>) Regularization parameters

Instance Members

▸ fit(x1, x2)

Fit model.

fit(x1: Array<Array<number>>, x2: Array<Array<number>>)

Parameters

x1 (Array<Array<number>>) Left value

x2 (Array<Array<number>>) Right value

▸ predict(x)

Returns estimated values.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

LSDDCPD

LSDD for change point detection

new LSDDCPD(w: number, take: number?, lag: number?)

Parameters

w (number) Window size

take (number?) Take number

lag (number?) Lag

Instance Members

▸ predict(datas)

Returns anomaly degrees.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

LSIF

lib/model/lsif.js

least-squares importance fitting

new LSIF(sigma: Array<number>, lambda: Array<number>, fold: number, kernelNum: number)

Parameters

sigma (Array<number>) Sigmas of normal distribution

lambda (Array<number>) Regularization parameters

fold (number) Number of folds

kernelNum (number) Number of kernels

Instance Members

▸ fit(x1, x2)

lib/model/lsif.js

Fit model.

fit(x1: Array<Array<number>>, x2: Array<Array<number>>)

Parameters

x1 (Array<Array<number>>) Numerator data

x2 (Array<Array<number>>) Denominator data

▸ predict(x)

lib/model/lsif.js

Returns estimated values.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

LeastTrimmedSquaresRegression

lib/model/lts.js

Least trimmed squares

new LeastTrimmedSquaresRegression(h: number?)

Parameters

h (number? = 0.9) Sampling rate

Instance Members

▸ fit(x, y)

lib/model/lts.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

lib/model/lts.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

LTSA

lib/model/ltsa.js

Local Tangent Space Alignment

new LTSA(k: number, rd: (number | null)?)

Parameters

k (number) Number of neighborhoods

rd ((number | null)? = null) Reduced dimension

Instance Members

▸ predict(x)

lib/model/ltsa.js

Returns reduced values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<number>>: Predicted values

LVQCluster

Learning Vector Quantization clustering

new LVQCluster(k: number)

Parameters

k (number) Number of clusters

Instance Members

▸ fit(x, lr)

Fit model.

fit(x: Array<Array<number>>, lr: number)

Parameters

x (Array<Array<number>>) Training data

lr (number = 0.1) Learning rate

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

LVQClassifier

Learning Vector Quantization classifier

new LVQClassifier(type: (1 | 2 | 3))

Parameters

type ((1 | 2 | 3)) Type number

Instance Members

▸ fit(x, y, lr)

Fit model.

fit(x: Array<Array<number>>, y: Array<any>, lr: number)

Parameters

x (Array<Array<number>>) Training data

y (Array<any>) Target values

lr (number = 0.1) Learning rate

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<any>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

MacQueenKMeans

MacQueen's k-Means algorithm

new MacQueenKMeans(k: number)

Parameters

k (number) Number of clusters

Instance Members

▸ centroids

Centroids

centroids

Type: Array<Array<number>>

▸ fit(datas)

Fit model.

fit(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

MAD

lib/model/mad.js

Median Absolute Deviation

new MAD()

Instance Members

▸ fit(data)

lib/model/mad.js

Fit model.

fit(data: Array<Array<number>>)

Parameters

data (Array<Array<number>>) Training data

▸ predict(data)

lib/model/mad.js

Returns anomaly degrees.

predict(data: Array<Array<number>>): Array<number>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

MADALINE

Many Adaptive Linear Neuron model

new MADALINE(sizes: Array<number>, rule: (1 | 2 | 3)?, rate: number)

Parameters

sizes (Array<number>) Numbers of layers

rule ((1 | 2 | 3)? = 2) Rule

rate (number) Learning rate

Instance Members

▸ fit(x, y)

Fit this model once.

fit(x: Array<Array<number>>, y: (Array<Array<(1 | -1)>> | Array<(1 | -1)>))

Parameters

x (Array<Array<number>>) Training data

y ((Array<Array<(1 | -1)>> | Array<(1 | -1)>)) Target values

▸ outputLayers(data, from)

Returns predicted datas.

outputLayers(data: Array<Array<number>>, from: number): Array<Array<Array<number>>>

Parameters

data (Array<Array<number>>) Sample data

from (number = 0) Index of layers to calculate from

Returns

Array<Array<Array<number>>>: Predicted values for each layer

▸ predict(data)

lib/model/margin_perceptron.js

Returns predicted datas.

predict(data: Array<Array<number>>): Array<Array<(1 | -1)>>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<Array<(1 | -1)>>: Predicted values

MarginPerceptron

Margin Perceptron

new MarginPerceptron(rate: number)

Parameters

rate (number) Learning rate

Instance Members

▸ fit(x, y)

lib/model/margin_perceptron.js

Fit model.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

lib/model/margin_perceptron.js

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

MarkovSwitching

Markov switching

new MarkovSwitching(regime: number)

Parameters

regime (number) Number of regime

Instance Members

▸ fit(datas, eps, trial)

Fit model.

fit(datas: Array<Array<number>>, eps: number, trial: number)

Parameters

datas (Array<Array<number>>) Training data

eps (number) Parameter update range

trial (number) Trial count

▸ probability(datas)

Returns probabilities.

probability(datas: Array<Array<number>>): Array<Array<number>>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

▸ predict(datas)

Returns anomaly degrees.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

MaxAbsScaler

lib/model/maxabs.js

Max absolute scaler

new MaxAbsScaler()

Instance Members

▸ fit(x)

lib/model/maxabs.js

Fit model.

fit(x: (Array<number> | Array<Array<number>>))

Parameters

x ((Array<number> | Array<Array<number>>)) Training data

▸ predict(x)

lib/model/maxabs.js

Returns transformed values.

predict(x: (Array<number> | Array<Array<number>>)): (Array<number> | Array<Array<number>>)

Parameters

x ((Array<number> | Array<Array<number>>)) Sample data

Returns

(Array<number> | Array<Array<number>>): Predicted values

MaximumLikelihoodEstimator

Maximum likelihood estimator

new MaximumLikelihoodEstimator(distribution: "normal"?)

Parameters

distribution ("normal"? = 'normal') Distribution name

Instance Members

▸ fit(x)

Fit model.

fit(x: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

▸ probability(x)

Returns probability of the data.

probability(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

▸ predict(x)

Returns probability of the data.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

MCD

lib/model/mcd.js

Minimum Covariance Determinant

new MCD(datas: Array<Array<number>>, sampling_rate: number)

Parameters

datas (Array<Array<number>>) Training data

sampling_rate (number) Sampling rate

Instance Members

▸ fit()

lib/model/mcd.js

Fit model.

fit()

▸ predict(data)

lib/model/mcd.js

Returns anomaly degrees.

predict(data: Array<Array<number>>): Array<number>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

MixtureDiscriminant

Mixture discriminant analysis

new MixtureDiscriminant(r: number)

Parameters

r (number) Number of components

Instance Members

▸ init(x, y)

Initialize this model.

init(x: Array<Array<number>>, y: Array<any>)

Parameters

x (Array<Array<number>>) Training data

y (Array<any>) Target values

▸ fit()

Fit model.

fit()

▸ predict(data)

Returns predicted categories.

predict(data: Array<Array<number>>): Array<any>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

MDS

lib/model/mds.js

Multi-dimensional Scaling

new MDS(rd: (number | null)?)

Parameters

rd ((number | null)? = null) Reduced dimension

Instance Members

▸ predict(x, dmat = false)

lib/model/mds.js

Returns reduced values.

predict(x: Array<Array<number>>, dmat: boolean?): Array<Array<number>>

Parameters

x (Array<Array<number>>) Training data

dmat (boolean? = false) True if the x is distance matrix.

Returns

Array<Array<number>>: Predicted values

MeanShift

Mean shift

new MeanShift(h: number, threshold: number)

Parameters

h (number) Smoothing parameter for the kernel

threshold (number) Threshold

Instance Members

▸ categories

Number of categories that last predicted

categories

Type: number

▸ h

h

Type: number

▸ init(data)

Initialize model.

init(data: Array<Array<number>>)

Parameters

data (Array<Array<number>>) Training data

▸ predict()

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

▸ fit()

Fit model.

fit(): boolean

Returns

boolean: true if any centroids has moved

MetropolisHastings

lib/model/mh.js

Metropolis-Hastings algorithm

new MetropolisHastings(targetFunc: function (Array<number>): number, d: number, q: "gaussian"?)

Parameters

targetFunc (function (Array<number>): number) Target distribution

d (number) Output size

q ("gaussian"? = 'gaussian') Proposal density name

Instance Members

▸ sample(n, t = 100)

lib/model/mh.js

Returns sampled values.

sample(n: number, t: number?): Array<Array<number>>

Parameters

n (number) Number of generated data

t (number? = 100) Iteration count for each generation

Returns

Array<Array<number>>: Generated values

MinmaxNormalization

lib/model/minmax.js

Min-max normalization

new MinmaxNormalization(min: number?, max: number?)

Parameters

min (number? = 0) Minimum value

max (number? = 1) Maximum value

Instance Members

▸ fit(x)

lib/model/minmax.js

Fit model.

fit(x: (Array<number> | Array<Array<number>>))

Parameters

x ((Array<number> | Array<Array<number>>)) Training data

▸ predict(x)

lib/model/minmax.js

Returns transformed values.

predict(x: (Array<number> | Array<Array<number>>)): (Array<number> | Array<Array<number>>)

Parameters

x ((Array<number> | Array<Array<number>>)) Sample data

Returns

(Array<number> | Array<Array<number>>): Predicted values

MIRA

Margin Infused Relaxed Algorithm

new MIRA()

Instance Members

▸ update(x, y)

Update model parameters with one data.

update(x: Array<number>, y: (1 | -1))

Parameters

x (Array<number>) Training data

y ((1 | -1)) Target value

▸ fit(x, y)

Fit model parameters.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

Returns predicted datas.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

MLLE

lib/model/mlle.js

Modified Locally Linear Embedding

new MLLE(k: number, rd: (number | null)?)

Parameters

k (number) Number of neighborhoods

rd ((number | null)? = null) Reduced dimension

Instance Members

▸ predict(data)

lib/model/mlle.js

Returns reduced values.

predict(data: Array<Array<number>>): Array<Array<number>>

Parameters

data (Array<Array<number>>) Training data

Returns

Array<Array<number>>: Predicted values

MLPClassifier

Multi layer perceptron classifier

Parameters

hidden_sizes (Array<number>) Sizes of hidden layers

Instance Members

▸ categories

Category list

categories

Type: Array<any>

▸ epoch

Epoch

epoch

Type: number

▸ toObject()

Returns object representation.

toObject(): Array<LayerObject>

Returns

Array<LayerObject>: Object represented this neuralnetwork

▸ fit(train_x, train_y, iteration, rate = 0.001, batch = 0)

Fit model.

fit(train_x: Array<Array<number>>, train_y: Array<any>, iteration: number, rate: number?, batch: number?): number

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<any>) Target values

iteration (number) Iteration count

rate (number? = 0.001) Learning rate

batch (number? = 0) Batch size

Returns

number: Loss value

▸ probability(x)

Returns predicted probabilities.

probability(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

▸ predict(x)

Returns predicted values.

predict(x: Array<Array<number>>): Array<any>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

MLPRegressor

Multi layer perceptron regressor

Parameters

hidden_sizes (Array<number>) Sizes of hidden layers

Instance Members

▸ epoch

Epoch

epoch

Type: number

▸ toObject()

Returns object representation.

toObject(): Array<LayerObject>

Returns

Array<LayerObject>: Object represented this neuralnetwork

▸ fit(train_x, train_y, iteration, rate = 0.001, batch = 0)

Fit model.

fit(train_x: Array<Array<number>>, train_y: Array<Array<number>>, iteration: number, rate: number?, batch: number?): number

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<Array<number>>) Target values

iteration (number) Iteration count

rate (number? = 0.001) Learning rate

batch (number? = 0) Batch size

Returns

number: Loss value

▸ predict(x)

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

MOD

lib/model/mod.js

Method of Optimal Direction

new MOD(x: Array<Array<number>>, m: number, k: number?)

Parameters

x (Array<Array<number>>) Training data

m (number) Reduced dimension

k (number? = m) Sparsity parameter

Instance Members

▸ fit()

lib/model/mod.js

Fit model and returns reduced values.

fit(): Array<Array<number>>

Returns

Array<Array<number>>: Predicted values

▸ predict()

lib/model/mod.js

Returns reduced values.

predict(): Array<Array<number>>

Returns

Array<Array<number>>: Predicted values

MONA

MONothetic Analysis Clustering

new MONA()

Instance Members

▸ size

Number of clusters

size

Type: number

▸ init(datas)

Initialize model.

init(datas: Array<Array<(0 | 1)>>)

Parameters

datas (Array<Array<(0 | 1)>>) Training data

▸ fit()

Fit model.

fit()

▸ predict()

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

MonotheticClustering

Monothetic Clustering

new MonotheticClustering()

Instance Members

▸ size

Number of clusters

size

Type: number

▸ init(datas)

Initialize model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ fit()

Fit model.

fit()

▸ predict()

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

MCAgent

Monte Carlo agent

new MCAgent(env: RLEnvironmentBase, resolution: number?)

Parameters

env (RLEnvironmentBase) Environment

resolution (number? = 20) Resolution

Instance Members

▸ reset()

Reset agent.

reset()

▸ get_score()

Returns a score.

get_score(): Array<Array<Array<number>>>

Returns

Array<Array<Array<number>>>: Score values

▸ get_action(state, greedy_rate)

Returns a action.

get_action(state: Array<any>, greedy_rate: number): Array<any>

Parameters

state (Array<any>) Current states

greedy_rate (number = 0.5) Greedy rate

Returns

Array<any>: Action

▸ update(action, state, reward, done)

Update model.

update(action: Array<any>, state: Array<any>, reward: number, done: boolean)

Parameters

action (Array<any>) Action

state (Array<any>) Next state

reward (number) Reward

done (boolean) Done epoch or not

Mountain

Mountain method

new Mountain(r: number, alpha: number, beta: number)

Parameters

r (number) Resolution of grid

alpha (number) Tuning parameter

beta (number) Tuning parameter

Instance Members

▸ init(datas)

Initialize model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ fit()

Fit model.

fit()

▸ predict(data)

Returns predicted categories.

predict(data: Array<Array<number>>): Array<number>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

SimpleMovingAverage

Simple moving average

new SimpleMovingAverage(n: number)

Parameters

n (number) Window size

Instance Members

▸ predict(data)

Returns smoothed values.

predict(data: Array<number>): Array<number>

Parameters

data (Array<number>) Training data

Returns

Array<number>: Predicted values

LinearWeightedMovingAverage

Linear weighted moving average

new LinearWeightedMovingAverage(n: number)

Parameters

n (number) Window size

Instance Members

▸ predict(data)

Returns smoothed values.

predict(data: Array<number>): Array<number>

Parameters

data (Array<number>) Training data

Returns

Array<number>: Predicted values

TriangularMovingAverage

Triangular moving average

new TriangularMovingAverage(k: number)

Parameters

k (number) Window size

Instance Members

▸ predict(data)

lib/model/moving_median.js

Returns smoothed values.

predict(data: Array<number>): Array<number>

Parameters

data (Array<number>) Training data

Returns

Array<number>: Predicted values

MovingMedian

Moving median

new MovingMedian(n: number)

Parameters

n (number) Window size

Instance Members

▸ predict(data)

lib/model/moving_median.js

Returns smoothed values.

predict(data: Array<number>): Array<number>

Parameters

data (Array<number>) Training data

Returns

Array<number>: Predicted values

MT

lib/model/mt.js

Mahalanobis Taguchi method

new MT()

Instance Members

▸ fit(data)

lib/model/mt.js

Fit model.

fit(data: Array<Array<number>>)

Parameters

data (Array<Array<number>>) Training data

▸ predict(data)

lib/model/mt.js

Returns anomaly degrees.

predict(data: Array<Array<number>>): Array<number>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

MutualInformationFeatureSelection

lib/model/mutual_information.js

Mutual information feature selector

new MutualInformationFeatureSelection(k: number)

Parameters

k (number) Number of selected features

Instance Members

▸ fit(x, y)

lib/model/mutual_information.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

lib/model/mutual_information.js

Returns feature selected values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

MutualKNN

Mutual k-nearest-neighbor model

new MutualKNN(k: number?)

Parameters

k (number? = 5) Number of neighborhoods

Instance Members

▸ size

Number of clusters

size

Type: number

▸ fit(datas)

Fit model.

fit(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ predict()

lib/model/n_cubic_interpolation.js

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

NCubicInterpolation

n-cubic interpolation

new NCubicInterpolation()

Instance Members

▸ fit(values, grids)

lib/model/n_cubic_interpolation.js

Fit model parameters.

fit(values: Array<any>, grids: Array<Array<number>>)

Parameters

values (Array<any>) Training data. Nested number array

grids (Array<Array<number>>) Grid values for each axises

▸ predict(x)

lib/model/n_cubic_interpolation.js

Returns predicted interpolated values.

predict(x: Array<Array<number>>): Array<(number | null)>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<(number | null)>: Predicted values

NLinearInterpolation

lib/model/n_linear_interpolation.js

n-linear interpolation

new NLinearInterpolation()

Instance Members

▸ fit(values, grids)

lib/model/n_linear_interpolation.js

Fit model parameters.

fit(values: Array<any>, grids: Array<Array<number>>)

Parameters

values (Array<any>) Training data. Nested number array

grids (Array<Array<number>>) Grid values for each axises

▸ predict(x)

lib/model/n_linear_interpolation.js

Returns predicted interpolated values.

predict(x: Array<Array<number>>): Array<(number | null)>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<(number | null)>: Predicted values

NadarayaWatson

lib/model/nadaraya_watson.js

Nadaraya–Watson kernel regression

new NadarayaWatson(s: number?)

Parameters

s (number?) Sigmas of normal distribution

Instance Members

▸ fit(x, y)

lib/model/nadaraya_watson.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

lib/model/nadaraya_watson.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

NaiveBayes

Naive bayes

new NaiveBayes(distribution: "gaussian"?)

Parameters

distribution ("gaussian"? = 'gaussian') Distribution name

Instance Members

▸ fit(datas, labels)

Fit model.

fit(datas: Array<Array<number>>, labels: Array<any>)

Parameters

datas (Array<Array<number>>) Training data

labels (Array<any>) Target values

▸ probability(data)

Returns predicted probabilities.

probability(data: Array<Array<number>>): Array<Array<number>>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

▸ predict(data)

lib/model/naive_bayes_regression.js

Returns predicted categories.

predict(data: Array<Array<number>>): Array<any>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

NaiveBayesRegression

Naive bayes regression

new NaiveBayesRegression(categoryPositions: Array<boolean>)

Parameters

categoryPositions (Array<boolean>) Category column position

Instance Members

▸ fit(x, y)

lib/model/naive_bayes_regression.js

Fit model.

fit(x: Array<Array<any>>, y: Array<number>)

Parameters

x (Array<Array<any>>) Training data

y (Array<number>) Target values

▸ predict(x)

lib/model/naive_bayes_regression.js

Returns predicted values.

predict(x: Array<Array<any>>): Array<number>

Parameters

x (Array<Array<any>>) Sample data

Returns

Array<number>: Predicted values

NAROW

Narrow Adaptive Regularization Of Weights

new NAROW(b: number?)

Parameters

b (number? = 1) Tuning parameter

Instance Members

▸ init(train_x, train_y)

Initialize this model.

init(train_x: Array<Array<number>>, train_y: Array<(1 | -1)>)

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<(1 | -1)>) Target values

▸ update(x, y)

Update model parameters with one data.

update(x: Matrix, y: (1 | -1))

Parameters

x (Matrix) Training data

y ((1 | -1)) Target value

▸ fit()

Fit model parameters.

fit()

▸ predict(data)

lib/model/natural_neighbor_interpolation.js

Returns predicted datas.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

NaturalNeighborInterpolation

Natural neighbor interpolation

new NaturalNeighborInterpolation()

Instance Members

▸ fit(x, y)

lib/model/natural_neighbor_interpolation.js

Fit model.

fit(x: Array<Array<number>>, y: Array<number>)

Parameters

x (Array<Array<number>>) Training data

y (Array<number>) Target values

▸ predict(x)

lib/model/natural_neighbor_interpolation.js

Returns probabilities of the datas.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

NeighbourhoodComponentsAnalysis

Neighbourhood components analysis

new NeighbourhoodComponentsAnalysis(d: number?, lr: number?)

Parameters

d (number? = null) Reduced dimension

lr (number? = 0.1) Learning rate

Instance Members

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<number>)

Parameters

x (Array<Array<number>>) Training data

y (Array<number>) Target values

▸ importance()

Returns importances of the features.

importance(): Array<number>

Returns

Array<number>: Importances

▸ predict(x)

Returns reduced datas.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

NearestCentroid

Nearest centroid classifier

new NearestCentroid(metric: ("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)?)

Parameters

metric (("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)? = 'euclid') Metric name

Instance Members

▸ add(point, category)

Add a data.

add(point: Array<number>, category: any)

Parameters

point (Array<number>) Training data

category (any) Target value

▸ fit(datas, targets)

Add datas.

fit(datas: Array<Array<number>>, targets: Array<any>)

Parameters

datas (Array<Array<number>>) Training data

targets (Array<any>) Target values

▸ predict(datas)

lib/model/negation_naive_bayes.js

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<any>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

NegationNaiveBayes

Negation Naive bayes

new NegationNaiveBayes(distribution: "gaussian"?)

Parameters

distribution ("gaussian"? = 'gaussian') Distribution name

Instance Members

▸ fit(datas, labels)

lib/model/negation_naive_bayes.js

Fit model.

fit(datas: Array<Array<number>>, labels: Array<any>)

Parameters

datas (Array<Array<number>>) Training data

labels (Array<any>) Target values

▸ predict(data)

lib/model/negation_naive_bayes.js

Returns predicted categories.

predict(data: Array<Array<number>>): Array<any>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

NeuralGas

Neural gas model

new NeuralGas(l: number?, m: number?)

Parameters

l (number? = 1) Neughborhood range

m (number? = 0.99) Decreasing factor of l

Instance Members

▸ centroids

Centroids

centroids

Type: Array<Array<number>>

▸ size

Number of clusters.

size

Type: number

▸ add(datas)

Add a new cluster.

add(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Added centroid

▸ clear()

Clear all clusters.

clear()

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

▸ fit(datas)

Fit model and returns total distance the centroid has moved.

fit(datas: Array<Array<number>>): number

Parameters

datas (Array<Array<number>>) Training data

Returns

number: Total distance the centroid has moved

NeuralnetworkException

Exception for neuralnetwork class

new NeuralnetworkException(message: string, value: any)

Extends Error

Parameters

message (string) Error message

value (any) Some value

NeuralNetwork

Neuralnetwork

new NeuralNetwork(graph: ComputationalGraph, optimizer: ("sgd" | "adam" | "momentum" | "rmsprop")?)

Parameters

graph (ComputationalGraph) Graph of a network

optimizer (("sgd" | "adam" | "momentum" | "rmsprop")? = 'sgd') Optimizer of the network

Static Members

▸ fromObject(layers, loss?, optimizer = 'sgd')

Returns neuralnetwork.

fromObject(layers: Array<LayerObject>, loss: string?, optimizer: ("sgd" | "adam" | "momentum" | "rmsprop")?): NeuralNetwork

Parameters

layers (Array<LayerObject>) Network layers

loss (string?) Loss name

optimizer (("sgd" | "adam" | "momentum" | "rmsprop")? = 'sgd') Optimizer of the network

Returns

NeuralNetwork: Created Neuralnetwork

▸ fromONNX(buffer)

Load onnx model.

fromONNX(buffer: (Uint8Array | ArrayBuffer | File)): Promise<NeuralNetwork>

Parameters

buffer ((Uint8Array | ArrayBuffer | File)) File

Returns

Promise<NeuralNetwork>: Loaded NeuralNetwork

Instance Members

▸ copy()

Returns a copy of this.

copy(): NeuralNetwork

Returns

NeuralNetwork: Copied network

▸ toObject()

Returns object representation.

toObject(): Array<LayerObject>

Returns

Array<LayerObject>: Object represented this neuralnetwork

▸ calc(x, t?, out?, options = {})

Returns calculated values.

calc(x: any, t: Matrix?, out: Array<string>?, options: object?)

Parameters

x (any)

t (Matrix?) Supervised value

out (Array<string>?) Name of node from which to get output

options (object? = {}) Option

▸ grad(e?)

Returns gradient values.

grad(e: Matrix?): Matrix

Parameters

e (Matrix?) Input of gradient

Returns

Matrix: Output of gradient

▸ update(learning_rate)

Update model parameters.

update(learning_rate: number)

Parameters

learning_rate (number) Learning rate

▸ fit(x, t, epoch = 1, learning_rate = 0.1, batch_size = null, options = {})

Fit model.

fit(x: any, t: (Array<Array<number>> | Matrix), epoch: number?, learning_rate: number?, batch_size: number?, options: object?): Array<number>

Parameters

x (any)

t ((Array<Array<number>> | Matrix)) Target values

epoch (number? = 1) Iteration count

learning_rate (number? = 0.1) Learning rate

batch_size (number? = null) Batch size

options (object? = {}) Option

Returns

Array<number>: Loss value

▸ predict(x)

Returns predicted values.

predict(x: any): Array<Array<number>>

Parameters

x (any)

Returns

Array<Array<number>>: Predicted values

NeuralnetworkLayerException

Exception for neuralnetwork layer class

new NeuralnetworkLayerException(message: string, value: any)

Extends Error

Parameters

message (string) Error message

value (any) Some value

Layer

Neuralnetwork layer

new Layer(obj: object)

Parameters

obj (object) Config

Static Members

▸ fromObject(obj)

Returns layer from JSON.

fromObject(obj: any): Layer

Parameters

obj (any)

Returns

Layer: Layer

▸ registLayer(name?, cls?)

Regist layer class.

registLayer(name: string?, cls: Layer?)

Parameters

name (string?) Name of the layer

cls (Layer?) Layer class

Instance Members

▸ bind(values)

Bind pre-condition values.

bind(values: object)

Parameters

values (object) Binding object

Name	Description
values.supervisor `Matrix?`	Supervisor data
values.n `number`	Data count
values.rest `any`	Some other values

▸ calc(x)

Returns calculated values.

calc(x: ...(Matrix | Tensor)): (Matrix | Tensor | Array<(Matrix | Tensor)>)

Parameters

x (...(Matrix | Tensor)) Input values

Returns

(Matrix | Tensor | Array<(Matrix | Tensor)>): Output values

▸ grad(bo)

Returns gradient values.

grad(bo: ...(Matrix | Tensor)): (Matrix | Tensor | Array<(Matrix | Tensor)>)

Parameters

bo (...(Matrix | Tensor)) Input value of backpropagation

Returns

(Matrix | Tensor | Array<(Matrix | Tensor)>): Output value of backpropagation

▸ update(optimizer)

Update parameters.

update(optimizer: {lr: number, delta: (function (string, Matrix): Matrix)})

Parameters

optimizer ({lr: number, delta: (function (string, Matrix): Matrix)}) Optimizer for this layer

▸ toObject()

Returns object of this layer.

toObject()

FlowLayer

Base class for Flow-based generative model

new FlowLayer()

Extends Layer

Instance Members

▸ inverse(y)

Returns inverse values.

inverse(y: ...(Matrix | Tensor)): (Matrix | Tensor | Array<(Matrix | Tensor)>)

Parameters

y (...(Matrix | Tensor)) Input value of inverse calculation

Returns

(Matrix | Tensor | Array<(Matrix | Tensor)>): Output value of inverse calculation

▸ jacobianDeterminant()

Returns determinant of the Jacobian.

jacobianDeterminant(): number

Returns

number: Determinant of the Jacobian

Node

Node

Type: object

Properties

layer (Layer) : Layer

name (string) : Name of the node

input (Array<string>?) : Input node names

outputValue ((Matrix | Array<Matrix>)?) : Output value of this node to next layer

gradientValue (Array<Matrix>?) : Gradient value of this node from next layer

ComputationalGraph

Computational graph for Neuralnetwork structure

new ComputationalGraph()

Static Members

▸ fromObject(nodes)

Returns Graph.

fromObject(nodes: Array<LayerObject>): ComputationalGraph

Parameters

nodes (Array<LayerObject>) Array of object represented a graph

Returns

ComputationalGraph: Graph

▸ fromONNX(buffer)

Load onnx model.

fromONNX(buffer: (Uint8Array | ArrayBuffer | File)): Promise<ComputationalGraph>

Parameters

buffer ((Uint8Array | ArrayBuffer | File)) File

Returns

Promise<ComputationalGraph>: Loaded graph

Instance Members

▸ nodes

Graph nodes

nodes

Type: Array<Node>

▸ inputNodes

Input nodes

inputNodes

Type: Array<Node>

▸ outputNodes

Output nodes

outputNodes

Type: Array<Node>

▸ size

Number of nodes

size

Type: number

▸ toObject()

Returns object representation.

toObject(): Array<LayerObject>

Returns

Array<LayerObject>: Object represented this graph

▸ toDot()

Returns a string of DOT format.

toDot(): string

Returns

string: String of DOT format

▸ add(layer, name?, inputs = undefined)

Add a layer.

add(layer: Layer, name: string?, inputs: (Array<string> | string)?)

Parameters

layer (Layer) Added layer

name (string?) Node name

inputs ((Array<string> | string)? = undefined) Input node names for the added layer

▸ bind(values)

Bind values to layers

bind(values: object)

Parameters

values (object) Binding values

▸ calc(require?)

Returns calculated values.

calc(require: Array<(string | number)>?)

Parameters

require (Array<(string | number)>?) Name or index of nodes at least calculated

▸ grad(e?)

Returns gradient values.

grad(e: Matrix?): Matrix

Parameters

e (Matrix?) Input of gradient

Returns

Matrix: Output of gradient

▸ getNode(name)

lib/model/nns/layer/index.js

Returns a specific name node.

getNode(name: string)

Parameters

name (string) Node name

AdditiveCoupling

lib/model/nns/layer/additive_coupling.js

Additive coupling layer

new AdditiveCoupling($0: Object, config: object)

Extends FlowLayer

Parameters

$0 (Object)

Name	Description
$0.d `any` (default `null`)
$0.net `any` (default `null`)
$0.rest `...any`

config (object) object

AdaptivePiecewiseLinearLayer

lib/model/nns/layer/apl.js

Adaptive piecewise linear layer

new AdaptivePiecewiseLinearLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.s `any` (default `2`)
$0.a `any` (default `0.1`)
$0.b `any` (default `0`)
$0.rest `...any`

config (object) object

ArandaLayer

lib/model/nns/layer/aranda.js

Aranda layer

new ArandaLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.l `any` (default `2`)
$0.rest `...any`

config (object) object

ArgmaxLayer

lib/model/nns/layer/argmax.js

Argmax layer

new ArgmaxLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.axis `any` (default `-1`)
$0.keepdims `any` (default `true`)
$0.rest `...any`

config (object) object

ArgminLayer

lib/model/nns/layer/argmin.js

Argmin layer

new ArgminLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.axis `any` (default `-1`)
$0.keepdims `any` (default `true`)
$0.rest `...any`

config (object) object

AttentionLayer

lib/model/nns/layer/attention.js

Attention layer

new AttentionLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.dk `any` (default `null`)
$0.dv `any` (default `null`)
$0.wq `any` (default `null`)
$0.wk `any` (default `null`)
$0.wv `any` (default `null`)
$0.rest `...any`

config (object) object

AveragePoolLayer

lib/model/nns/layer/averagepool.js

Average pool layer

new AveragePoolLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.kernel `any`
$0.stride `any` (default `null`)
$0.padding `any` (default `null`)
$0.channel_dim `any` (default `-1`)
$0.rest `...any`

config (object) object

BatchNormalizationLayer

lib/model/nns/layer/batch_normalization.js

Batch normalization layer

new BatchNormalizationLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.scale `any` (default `1`)
$0.offset `any` (default `0`)
$0.epsilon `any` (default `1.0e-12`)
$0.channel_dim `any` (default `-1`)
$0.input_mean `any`
$0.input_var `any`
$0.rest `...any`

config (object) object

BimodalDerivativeAdaptiveActivationLayer

lib/model/nns/layer/bdaa.js

Bimodal derivative adaptive activation layer

new BimodalDerivativeAdaptiveActivationLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.alpha `any` (default `1`)
$0.rest `...any`

config (object) object

BendableLinearUnitLayer

lib/model/nns/layer/blu.js

Bendable linear unit layer

new BendableLinearUnitLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.beta `any` (default `0.1`)
$0.rest `...any`

config (object) object

BoundedReLULayer

lib/model/nns/layer/brelu.js

Bounded ReLU layer

new BoundedReLULayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.a `any` (default `1`)
$0.rest `...any`

config (object) object

ContinuouslyDifferentiableELULayer

lib/model/nns/layer/celu.js

Continuously differentiable ELU layer

new ContinuouslyDifferentiableELULayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.a `any` (default `1.0`)
$0.rest `...any`

config (object) object

ClipLayer

lib/model/nns/layer/clip.js

Clip layer

new ClipLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.min `any` (default `null`)
$0.max `any` (default `null`)
$0.rest `...any`

config (object) object

ConcatLayer

lib/model/nns/layer/concat.js

Concat layer

new ConcatLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.axis `any` (default `1`)
$0.rest `...any`

config (object) object

CondLayer

lib/model/nns/layer/cond.js

Condition layer

new CondLayer()

Extends Layer

ConstLayer

lib/model/nns/layer/const.js

Constant layer

new ConstLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.value `any`
$0.rest `...any`

config (object) object

ConvLayer

lib/model/nns/layer/conv.js

Convolutional layer

new ConvLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.kernel `any`
$0.channel `any` (default `null`)
$0.stride `any` (default `null`)
$0.padding `any` (default `null`)
$0.w `any` (default `null`)
$0.activation `any` (default `null`)
$0.l2_decay `any` (default `0`)
$0.l1_decay `any` (default `0`)
$0.channel_dim `any` (default `-1`)
$0.rest `...any`

config (object) object

ConcatenatedReLULayer

lib/model/nns/layer/crelu.js

Concatenated ReLU layer

new ConcatenatedReLULayer()

Extends Layer

DropoutLayer

lib/model/nns/layer/dropout.js

Dropout layer

new DropoutLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.drop_rate `any` (default `0.5`)
$0.rest `...any`

config (object) object

ElasticELULayer

lib/model/nns/layer/eelu.js

Elastic ELU layer

new ElasticELULayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.k `any` (default `1`)
$0.alpha `any` (default `1`)
$0.beta `any` (default `1`)
$0.rest `...any`

config (object) object

ELULayer

lib/model/nns/layer/elu.js

ELU layer

new ELULayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.a `any` (default `1`)
$0.rest `...any`

config (object) object

EmbeddingLayer

lib/model/nns/layer/embedding.js

Embedding layer

new EmbeddingLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.size `any` (default `512`)
$0.embeddings `any` (default `{}`)
$0.rest `...any`

config (object) object

ElasticReLULayer

lib/model/nns/layer/erelu.js

Elastic ReLU layer

new ElasticReLULayer($0: Object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.rest `...any`

ESwishLayer

lib/model/nns/layer/eswish.js

E-swish layer

new ESwishLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.beta `any` (default `1`)
$0.rest `...any`

config (object) object

FastELULayer

lib/model/nns/layer/felu.js

Fast ELU layer

new FastELULayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.alpha `any` (default `1`)
$0.rest `...any`

config (object) object

FlattenLayer

lib/model/nns/layer/flatten.js

Flatten layer

new FlattenLayer()

Extends Layer

FlexibleReLULayer

lib/model/nns/layer/frelu.js

Flexible ReLU layer

new FlexibleReLULayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.b `any` (default `0`)
$0.rest `...any`

config (object) object

FullyConnected

lib/model/nns/layer/full.js

Fully connected layer

new FullyConnected($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.out_size `any`
$0.w `any` (default `null`)
$0.b `any` (default `null`)
$0.activation `any` (default `null`)
$0.l2_decay `any` (default `0`)
$0.l1_decay `any` (default `0`)
$0.rest `...any`

config (object) object

FunctionLayer

lib/model/nns/layer/function.js

Function layer

new FunctionLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.func `any`
$0.rest `...any`

config (object) config

GaussianLayer

lib/model/nns/layer/gaussian.js

Gaussian layer

new GaussianLayer()

Extends Layer

GlobalAveragePoolLayer

lib/model/nns/layer/global_averagepool.js

Global average pool layer

new GlobalAveragePoolLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.channel_dim `any` (default `-1`)
$0.rest `...any`

config (object) object

GlobalLpPoolLayer

lib/model/nns/layer/global_lppool.js

Global Lp pool layer

new GlobalLpPoolLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.p `any` (default `2`)
$0.channel_dim `any` (default `-1`)
$0.rest `...any`

config (object) object

GlobalMaxPoolLayer

lib/model/nns/layer/global_maxpool.js

Global max pool layer

new GlobalMaxPoolLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.channel_dim `any` (default `-1`)
$0.rest `...any`

config (object) object

GraphConvolutionalLayer

lib/model/nns/layer/graph_conv.js

Graph convolutional layer

new GraphConvolutionalLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.out_size `any`
$0.w `any` (default `null`)
$0.b `any` (default `null`)
$0.activation `any` (default `null`)
$0.l2_decay `any` (default `0`)
$0.l1_decay `any` (default `0`)
$0.rest `...any`

config (object) object

GraphSAGELayer

lib/model/nns/layer/graph_sage.js

Graph SAGE layer

new GraphSAGELayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.out_size `any`
$0.aggregate `any` (default `'mean'`)
$0.w `any` (default `null`)
$0.b `any` (default `null`)
$0.activation `any` (default `null`)
$0.l2_decay `any` (default `0`)
$0.l1_decay `any` (default `0`)
$0.rest `...any`

config (object) object

GRULayer

lib/model/nns/layer/gru.js

GRU layer

new GRULayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.size `any`
$0.return_sequences `any` (default `false`)
$0.w_z `any` (default `null`)
$0.w_r `any` (default `null`)
$0.w_h `any` (default `null`)
$0.u_z `any` (default `null`)
$0.u_r `any` (default `null`)
$0.u_h `any` (default `null`)
$0.b_z `any` (default `null`)
$0.b_r `any` (default `null`)
$0.b_h `any` (default `null`)
$0.sequence_dim `any` (default `1`)
$0.rest `...any`

config (object) object

HardShrinkLayer

lib/model/nns/layer/hard_shrink.js

Hard shrink layer

new HardShrinkLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.l `any` (default `0.5`)
$0.rest `...any`

config (object) object

HardSigmoidLayer

lib/model/nns/layer/hard_sigmoid.js

Hard sigmoid layer

new HardSigmoidLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.alpha `any` (default `0.2`)
$0.beta `any` (default `0.5`)
$0.rest `...any`

config (object) object

HardTanhLayer

lib/model/nns/layer/hard_tanh.js

Hard tanh layer

new HardTanhLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.v `any` (default `1`)
$0.rest `...any`

config (object) object

HexpoLayer

lib/model/nns/layer/hexpo.js

Hexpo layer

new HexpoLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.a `any` (default `1`)
$0.b `any` (default `1`)
$0.c `any` (default `1`)
$0.d `any` (default `1`)
$0.rest `...any`

config (object) object

HuberLayer

lib/model/nns/layer/huber.js

Huber loss layer

new HuberLayer()

Extends Layer

IncludeLayer

lib/model/nns/layer/include.js

Include layer

new IncludeLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.net `any`
$0.input_to `any` (default `null`)
$0.train `any` (default `true`)
$0.rest `...any`

config (object) object

InputLayer

lib/model/nns/layer/input.js

Input layer

new InputLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.name `any` (default `null`)
$0.rest `...any`

config (object) object

ImprovedSigmoidLayer

lib/model/nns/layer/isigmoid.js

Improved sigmoid layer

new ImprovedSigmoidLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.a `any` (default `1`)
$0.alpha `any` (default `1`)
$0.rest `...any`

config (object) object

LayerNormalizationLayer

lib/model/nns/layer/layer_normalization.js

Layer normalization layer

new LayerNormalizationLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.axis `any` (default `-1`)
$0.epsilon `any` (default `1.0e-12`)
$0.scale `any` (default `1`)
$0.offset `any` (default `0`)
$0.rest `...any`

config (object) object

LeakyReLULayer

lib/model/nns/layer/leaky_relu.js

Leaky ReLU layer

new LeakyReLULayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.a `any` (default `0.1`)
$0.rest `...any`

config (object) config

LogSoftmaxLayer

lib/model/nns/layer/logsoftmax.js

Log softmax layer

new LogSoftmaxLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.axis `any` (default `-1`)
$0.rest `...any`

config (object) object

LpPoolLayer

lib/model/nns/layer/lppool.js

Lp pool layer

new LpPoolLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.p `any` (default `2`)
$0.kernel `any`
$0.stride `any` (default `null`)
$0.padding `any` (default `null`)
$0.channel_dim `any` (default `-1`)
$0.rest `...any`

config (object) object

LRNLayer

lib/model/nns/layer/lrn.js

LRN layer

new LRNLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.alpha `any` (default `0.0001`)
$0.beta `any` (default `0.75`)
$0.k `any` (default `2`)
$0.n `any`
$0.channel_dim `any` (default `-1`)
$0.rest `...any`

config (object) object

LSTMLayer

lib/model/nns/layer/lstm.js

LSTM layer

new LSTMLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.size `any`
$0.return_sequences `any` (default `false`)
$0.w_z `any` (default `null`)
$0.w_in `any` (default `null`)
$0.w_for `any` (default `null`)
$0.w_out `any` (default `null`)
$0.r_z `any` (default `null`)
$0.r_in `any` (default `null`)
$0.r_for `any` (default `null`)
$0.r_out `any` (default `null`)
$0.p_in `any` (default `null`)
$0.p_for `any` (default `null`)
$0.p_out `any` (default `null`)
$0.b_z `any` (default `null`)
$0.b_in `any` (default `null`)
$0.b_for `any` (default `null`)
$0.b_out `any` (default `null`)
$0.sequence_dim `any` (default `1`)
$0.rest `...any`

config (object) object

MatmulLayer

lib/model/nns/layer/matmul.js

Matrix multiply layer

new MatmulLayer()

Extends Layer

MaxPoolLayer

lib/model/nns/layer/maxpool.js

Max pool layer

new MaxPoolLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.kernel `any`
$0.stride `any` (default `null`)
$0.padding `any` (default `null`)
$0.channel_dim `any` (default `-1`)
$0.rest `...any`

config (object) object

MeanLayer

lib/model/nns/layer/mean.js

Reduce mean layer

new MeanLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.axis `any` (default `-1`)
$0.keepdims `any` (default `true`)
$0.rest `...any`

config (object) config

MultipleParametricELULayer

lib/model/nns/layer/mpelu.js

Multiple parametric ELU layer

new MultipleParametricELULayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.alpha `any` (default `1`)
$0.beta `any` (default `1`)
$0.rest `...any`

config (object) config

MSELayer

lib/model/nns/layer/mse.js

MSE loss layer

new MSELayer()

Extends Layer

MultibinTrainableLinearUnitLayer

lib/model/nns/layer/mtlu.js

Multibin trainable linear unit layer

new MultibinTrainableLinearUnitLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.a `any` (default `1`)
$0.b `any` (default `0`)
$0.c `any` (default `null`)
$0.k `any` (default `10`)
$0.rest `...any`

config (object) config

NaturalLogarithmReLULayer

lib/model/nns/layer/nlrelu.js

Natural logarithm ReLU layer

new NaturalLogarithmReLULayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.beta `any` (default `1`)
$0.rest `...any`

config (object) config

OnehotLayer

lib/model/nns/layer/onehot.js

One-hot layer

new OnehotLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.class_size `any` (default `null`)
$0.values `any` (default `[]`)
$0.rest `...any`

config (object) config

OutputLayer

lib/model/nns/layer/output.js

Output layer

new OutputLayer()

Extends Layer

PadeActivationUnitLayer

lib/model/nns/layer/pau.js

Pade activation unit layer

new PadeActivationUnitLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.m `any` (default `2`)
$0.n `any` (default `2`)
$0.a `any` (default `0.1`)
$0.b `any` (default `0`)
$0.rest `...any`

config (object) config

ParametricDeformableELULayer

lib/model/nns/layer/pdelu.js

Parametric deformable ELU layer

new ParametricDeformableELULayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.t `any` (default `0.1`)
$0.alpha `any` (default `1`)
$0.rest `...any`

config (object) config

ParametricELULayer

lib/model/nns/layer/pelu.js

Parametric ELU layer

new ParametricELULayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.a `any` (default `1`)
$0.b `any` (default `1`)
$0.rest `...any`

config (object) config

PiecewiseLinearUnitLayer

lib/model/nns/layer/plu.js

Piecewise linear unit layer

new PiecewiseLinearUnitLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.alpha `any` (default `0.1`)
$0.c `any` (default `1`)
$0.rest `...any`

config (object) config

ParametricReLULayer

lib/model/nns/layer/prelu.js

Parametric ReLU layer

new ParametricReLULayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.a `any` (default `0.25`)
$0.rest `...any`

config (object) config

ParametricRectifiedExponentialUnitLayer

lib/model/nns/layer/preu.js

Parametric rectified exponential unit layer

new ParametricRectifiedExponentialUnitLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.alpha `any` (default `1`)
$0.beta `any` (default `1`)
$0.rest `...any`

config (object) config

ProdLayer

lib/model/nns/layer/prod.js

Reduce product layer

new ProdLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.axis `any` (default `-1`)
$0.keepdims `any` (default `true`)
$0.rest `...any`

config (object) config

ParametricSigmoidFunctionLayer

lib/model/nns/layer/psf.js

Parametric sigmoid function layer

new ParametricSigmoidFunctionLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.m `any` (default `2`)
$0.rest `...any`

config (object) config

PenalizedTanhLayer

lib/model/nns/layer/ptanh.js

Penalized tanh layer

new PenalizedTanhLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.a `any` (default `0.25`)
$0.rest `...any`

config (object) config

ParametricTanhLinearUnitLayer

lib/model/nns/layer/ptelu.js

Parametric tanh linear unit layer

new ParametricTanhLinearUnitLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.alpha `any` (default `1`)
$0.beta `any` (default `1`)
$0.rest `...any`

config (object) config

RandomLayer

lib/model/nns/layer/random.js

Random layer

new RandomLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.size `any`
$0.mean `any` (default `0`)
$0.variance `any` (default `1`)
$0.rest `...any`

config (object) config

ReadoutLayer

lib/model/nns/layer/readout.js

Readout layer

new ReadoutLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.method `any` (default `'mean'`)
$0.rest `...any`

config (object) config

ReduceMaxLayer

lib/model/nns/layer/reduce_max.js

Reduce max layer

new ReduceMaxLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.axis `any` (default `-1`)
$0.keepdims `any` (default `true`)
$0.rest `...any`

config (object) config

ReduceMinLayer

lib/model/nns/layer/reduce_min.js

Reduce min layer

new ReduceMinLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.axis `any` (default `-1`)
$0.keepdims `any` (default `true`)
$0.rest `...any`

config (object) config

RectifiedPowerUnitLayer

lib/model/nns/layer/repu.js

Rectified power unit layer

new RectifiedPowerUnitLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.s `any` (default `2`)
$0.rest `...any`

config (object) config

ReshapeLayer

lib/model/nns/layer/reshape.js

Reshape layer

new ReshapeLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.size `any`
$0.rest `...any`

config (object) config

RNNLayer

lib/model/nns/layer/rnn.js

Simple RNN layer

new RNNLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.size `any`
$0.activation `any` (default `'tanh'`)
$0.return_sequences `any` (default `false`)
$0.w_x `any` (default `null`)
$0.w_h `any` (default `null`)
$0.b_x `any` (default `null`)
$0.b_h `any` (default `null`)
$0.sequence_dim `any` (default `1`)
$0.rest `...any`

config (object) object

RandomizedReLULayer

lib/model/nns/layer/rrelu.js

Randomized ReLU layer

new RandomizedReLULayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.l `any` (default `1.0/8`)
$0.u `any` (default `1.0/3`)
$0.rest `...any`

config (object) config

RandomTranslationReLULayer

lib/model/nns/layer/rtrelu.js

Random translation ReLU layer

new RandomTranslationReLULayer($0: Object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.rest `...any`

ScaledELULayer

lib/model/nns/layer/selu.js

Scaled ELU layer

new ScaledELULayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.a `any` (default `1.67326319217681884765625`)
$0.g `any` (default `1.05070102214813232421875`)
$0.rest `...any`

config (object) config

SigmoidLayer

lib/model/nns/layer/sigmoid.js

Sigmoid layer

new SigmoidLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.a `any` (default `1`)
$0.rest `...any`

config (object) config

SelfLearnableAFLayer

lib/model/nns/layer/slaf.js

Self learnable AF layer

new SelfLearnableAFLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.n `any` (default `3`)
$0.a `any` (default `1`)
$0.rest `...any`

config (object) config

SoftplusLinearUnitLayer

lib/model/nns/layer/slu.js

Softplus linear unit layer

new SoftplusLinearUnitLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.alpha `any` (default `1`)
$0.beta `any` (default `1`)
$0.gamma `any` (default `0`)
$0.rest `...any`

config (object) config

SoftShrinkLayer

lib/model/nns/layer/soft_shrink.js

Soft shrink layer

new SoftShrinkLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.l `any` (default `0.5`)
$0.rest `...any`

config (object) config

SoftargmaxLayer

lib/model/nns/layer/softargmax.js

Softargmax layer

new SoftargmaxLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.beta `any` (default `10000`)
$0.rest `...any`

config (object) config

SoftmaxLayer

lib/model/nns/layer/softmax.js

Softmax layer

new SoftmaxLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.axis `any` (default `-1`)
$0.rest `...any`

config (object) object

SoftminLayer

lib/model/nns/layer/softmin.js

Softmin layer

new SoftminLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.axis `any` (default `-1`)
$0.rest `...any`

config (object) object

SoftplusLayer

lib/model/nns/layer/softplus.js

Softplus layer

new SoftplusLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.beta `any` (default `1`)
$0.rest `...any`

config (object) config

SparseLayer

lib/model/nns/layer/sparse.js

Sparse layer

new SparseLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.rho `any`
$0.beta `any`
$0.rest `...any`

config (object) config

SplitLayer

lib/model/nns/layer/split.js

Split layer

new SplitLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.axis `any` (default `1`)
$0.size `any`
$0.rest `...any`

config (object) config

ShiftedReLULayer

lib/model/nns/layer/srelu.js

Shifted ReLU layer

new ShiftedReLULayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.d `any` (default `0`)
$0.rest `...any`

config (object) config

SoftRootSignLayer

lib/model/nns/layer/srs.js

Soft root sign layer

new SoftRootSignLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.alpha `any` (default `3`)
$0.beta `any` (default `2`)
$0.rest `...any`

config (object) config

ScaledTanhLayer

lib/model/nns/layer/stanh.js

Scaled tanh layer

new ScaledTanhLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.a `any` (default `1`)
$0.b `any` (default `1`)
$0.rest `...any`

config (object) config

StdLayer

lib/model/nns/layer/std.js

Standard deviation layer

new StdLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.axis `any` (default `-1`)
$0.keepdims `any` (default `true`)
$0.rest `...any`

config (object) config

SumLayer

lib/model/nns/layer/sum.js

Reduce sum layer

new SumLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.axis `any` (default `-1`)
$0.keepdims `any` (default `true`)
$0.rest `...any`

config (object) config

SupervisorLayer

lib/model/nns/layer/supervisor.js

Supervisor layer

new SupervisorLayer()

Extends Layer

SwishLayer

lib/model/nns/layer/swish.js

Swish layer

new SwishLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.beta `any` (default `1`)
$0.rest `...any`

config (object) config

TrainableAFLayer

lib/model/nns/layer/taf.js

Trainable AF layer

new TrainableAFLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.a `any` (default `0`)
$0.b `any` (default `0`)
$0.rest `...any`

config (object) config

ThresholdedReLULayer

lib/model/nns/layer/thresholded_relu.js

Thresholded ReLU layer

new ThresholdedReLULayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.a `any` (default `1.0`)
$0.rest `...any`

config (object) config

TransposeLayer

lib/model/nns/layer/transpose.js

Transpose layer

new TransposeLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.axis `any`
$0.rest `...any`

config (object) config

VariableLayer

lib/model/nns/layer/variable.js

Variable layer

new VariableLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.size `any`
$0.l2_decay `any` (default `0`)
$0.l1_decay `any` (default `0`)
$0.value `any` (default `null`)
$0.rest `...any`

config (object) config

VarLayer

lib/model/nns/layer/variance.js

Variance layer

new VarLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.axis `any` (default `-1`)
$0.keepdims `any` (default `true`)
$0.rest `...any`

config (object) config

NiblackThresholding

lib/model/niblack.js

Niblack thresholding

new NiblackThresholding(n: number?, k: number?)

Parameters

n (number? = 3) Size of local range

k (number? = 0.1) Tuning parameter

Instance Members

▸ predict(x)

lib/model/niblack.js

Returns thresholded values.

predict(x: Array<Array<number>>): Array<Array<(0 | 1)>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<(0 | 1)>>: Predicted values

NICE

Flow-based generative model non-linear independent component estimation

new NICE(layer_number: number, optimizer: string)

Parameters

layer_number (number) Number of layers

optimizer (string) Optimizer of the network

Instance Members

▸ epoch

Epoch

epoch

Type: number

▸ fit(x, iteration, rate, batch_size)

Fit model.

fit(x: Array<Array<number>>, iteration: number, rate: number, batch_size: number)

Parameters

x (Array<Array<number>>) Training data

iteration (number) Iteration count

rate (number) Learning rate

batch_size (number) Batch size

▸ predict(x)

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

▸ generate(z)

lib/model/nns/layer/reverse.js

Returns generated values.

generate(z: Array<Array<number>>): Array<Array<number>>

Parameters

z (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Generated values

ReverseLayer

Reverse layer

new ReverseLayer($0: Object, config: object)

Extends Layer

Parameters

$0 (Object)

Name	Description
$0.axis `any` (default `1`)
$0.rest `...any`

config (object) config

NLMeans

lib/model/nlmeans.js

Non-local means filter

new NLMeans(n: number, h: number)

Parameters

n (number) Manhattan distance of the pixel to the nearest neighbor

h (number) Degree of filtering

Instance Members

▸ predict(x)

lib/model/nlmeans.js

Returns predicted values.

predict(x: Array<Array<Array<number>>>): Array<Array<Array<number>>>

Parameters

x (Array<Array<Array<number>>>) Training data

Returns

Array<Array<Array<number>>>: Predicted values

NMF

Non-negative matrix factorization

new NMF()

Instance Members

▸ init(x, rd = 0)

Initialize model.

init(x: Array<Array<number>>, rd: number?)

Parameters

x (Array<Array<number>>) Training data

rd (number? = 0) Reduced dimension

▸ fit()

Fit model.

fit()

▸ predict()

Returns reduced datas.

predict(): Array<Array<number>>

Returns

Array<Array<number>>: Predicted values

NNBCA

lib/model/nnbca.js

Natural Neighborhood Based Classification Algorithm

new NNBCA(metric: ("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)?)

Parameters

metric (("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)? = 'euclid') Metric name

Instance Members

▸ fit(datas, targets)

lib/model/nnbca.js

Add datas.

fit(datas: Array<Array<number>>, targets: Array<any>)

Parameters

datas (Array<Array<number>>) Training data

targets (Array<any>) Target values

▸ predict(datas)

lib/model/nnbca.js

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<(any | null)>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<(any | null)>: Predicted values

NOF

lib/model/nof.js

Natural Outlier Factor

new NOF(k: number?)

Parameters

k (number? = 0) Number of neighborhoods

Instance Members

▸ predict(datas)

lib/model/nof.js

Returns anomaly degrees.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Predicted values

NormalHERD

Normal Herd

new NormalHERD(type: ("full" | "exact" | "project" | "drop")?, c: number?)

Parameters

type (("full" | "exact" | "project" | "drop")? = 'exact') Method name

c (number? = 0.1) Tradeoff value between passiveness and aggressiveness

Instance Members

▸ init(train_x, train_y)

Initialize this model.

init(train_x: Array<Array<number>>, train_y: Array<(1 | -1)>)

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<(1 | -1)>) Target values

▸ update(x, y)

Update model parameters with one data.

update(x: Matrix, y: (1 | -1))

Parameters

x (Matrix) Training data

y ((1 | -1)) Target value

▸ fit()

Fit model parameters.

fit()

▸ predict(data)

Returns predicted datas.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

OAPBPM

lib/model/oapbpm.js

Online Aggregate Prank-Bayes Point Machine

new OAPBPM(n: number, tau: number, rate: number?)

Parameters

n (number) Number of PRank models

tau (number) Probability to learn

rate (number? = 0.1) Learning rate

Instance Members

▸ fit(x, y)

lib/model/oapbpm.js

Fit model.

fit(x: Array<Array<number>>, y: Array<number>)

Parameters

x (Array<Array<number>>) Training data

y (Array<number>) Target values

▸ predict(x)

lib/model/oapbpm.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

OCSVM

One-class support vector machine

new OCSVM(nu: number, kernel: any)

Parameters

nu (number) Nu

kernel (any)

Instance Members

▸ init(x)

Initialize model.

init(x: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

▸ fit()

Fit model.

fit()

▸ predict(x)

Returns anomaly degrees.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

ODIN

lib/model/odin.js

Outlier Detection using Indegree Number

new ODIN(k: number?, t: number?)

Parameters

k (number? = 5) Number of neighborhoods

t (number? = 0) Indegree threshold

Instance Members

▸ predict(data)

lib/model/odin.js

Returns a list of the data predicted as outliers or not.

predict(data: Array<Array<number>>): Array<boolean>

Parameters

data (Array<Array<number>>) Training data

Returns

Array<boolean>: Predicted values

OnlineGradientDescent

Online gradient descent

new OnlineGradientDescent(c: number?, loss: "zero_one"?)

Parameters

c (number? = 1) Tuning parameter

loss ("zero_one"? = 'zero_one') Loss type name

Instance Members

▸ update(x, y)

Update model parameters with one data.

update(x: Array<number>, y: (1 | -1))

Parameters

x (Array<number>) Training data

y ((1 | -1)) Target value

▸ fit(x, y)

Fit model parameters.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

lib/model/ordered_logistic.js

Returns predicted datas.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

OPTICS

lib/model/optics.js

Ordering points to identify the clustering structure

new OPTICS(threshold: number, eps: number?, minPts: number?, metric: ("euclid" | "manhattan" | "chebyshev" | function (Array<number>, Array<number>): number)?)

Parameters

threshold (number) Threshold

eps (number? = Infinity) Radius to determine neighborhood

minPts (number? = 5) Number of neighborhood with core distance

metric (("euclid" | "manhattan" | "chebyshev" | function (Array<number>, Array<number>): number)? = 'euclid') Metric name

Instance Members

▸ fit(datas)

lib/model/optics.js

Fit model.

fit(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ predict()

lib/model/optics.js

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

ORCLUS

lib/model/orclus.js

arbitrarily ORiented projected CLUSter generation

new ORCLUS(k: number, k0: number, l: number)

Parameters

k (number) Number of clusters

k0 (number) Number of begining seeds

l (number) Number of dimensions

Instance Members

▸ fit(datas)

lib/model/orclus.js

Fit model.

fit(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ predict(datas)

lib/model/orclus.js

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

OrderedLogisticRegression

Ordered logistic regression

new OrderedLogisticRegression(rate: number?)

Parameters

rate (number? = 0.1) Learning rate

Instance Members

▸ fit(x, y)

lib/model/ordered_logistic.js

Fit model.

fit(x: Array<Array<number>>, y: Array<number>)

Parameters

x (Array<Array<number>>) Training data

y (Array<number>) Target values

▸ predict(x)

lib/model/ordered_logistic.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

OrderedProbitRegression

lib/model/ordered_probit.js

Ordered probit regression

new OrderedProbitRegression(rate: number?)

Parameters

rate (number? = 0.001) Learning rate

Instance Members

▸ fit(x, y)

lib/model/ordered_probit.js

Fit model.

fit(x: Array<Array<number>>, y: Array<number>)

Parameters

x (Array<Array<number>>) Training data

y (Array<number>) Target values

▸ predict(x)

lib/model/ordered_probit.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

OtsusThresholding

lib/model/otsu.js

Otus's thresholding

new OtsusThresholding()

Instance Members

▸ predict(x)

lib/model/otsu.js

Returns thresholded values.

predict(x: Array<number>): Array<(0 | 1)>

Parameters

x (Array<number>) Training data

Returns

Array<(0 | 1)>: Predicted values

PAM

Partitioning Around Medoids

new PAM(k: number)

Parameters

k (number) Number of clusters

Instance Members

▸ init(datas)

Initialize model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ fit()

Fit model and returns true if any centroids has moved.

fit(): boolean

Returns

boolean: true if any centroids has moved

▸ predict()

lib/model/particle_filter.js

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

ParticleFilter

Particle filter

new ParticleFilter()

Instance Members

▸ fit(z)

lib/model/particle_filter.js

Fit and returns smoothed values.

fit(z: Array<Array<number>>): Array<Array<number>>

Parameters

z (Array<Array<number>>) Training data

Returns

Array<Array<number>>: Predicted values

PassingBablok

lib/model/passing_bablok.js

Passing-Bablok method

new PassingBablok()

Instance Members

▸ fit(x, y)

lib/model/passing_bablok.js

Fit model.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(x)

lib/model/passing_bablok.js

Returns predicted values.

predict(x: Array<number>): Array<number>

Parameters

x (Array<number>) Sample data

Returns

Array<number>: Predicted values

PA

Passive Aggressive

new PA(v: (0 | 1 | 2)?)

Parameters

v ((0 | 1 | 2)? = 0) Version number

Instance Members

▸ update(x, y)

Update model parameters with one data.

update(x: Array<number>, y: (1 | -1))

Parameters

x (Array<number>) Training data

y ((1 | -1)) Target value

▸ fit(x, y)

Fit model parameters.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

Returns predicted datas.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

PAUM

lib/model/paum.js

Perceptron Algorithm with Uneven Margins

new PAUM(rate: number, tp: number, tm: number)

Parameters

rate (number) Learning rate

tp (number) Margin parameter for +1

tm (number) Margin parameter for -1

Instance Members

▸ fit(x, y)

lib/model/paum.js

Fit model.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

lib/model/paum.js

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

PCA

Principal component analysis

new PCA(rd: (number | null)?)

Parameters

rd ((number | null)? = null) Reduced dimension

Instance Members

▸ fit(x)

Fit model.

fit(x: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

▸ predict(x)

Returns reduced datas.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

DualPCA

Dual Principal component analysis

new DualPCA(rd: (number | null)?)

Parameters

rd ((number | null)? = null) Reduced dimension

Instance Members

▸ fit(x)

Fit model.

fit(x: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

▸ predict(x)

Returns reduced datas.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

KernelPCA

Kernel Principal component analysis

new KernelPCA(kernel: any, rd: (number | null)?)

Parameters

kernel (any)

rd ((number | null)? = null) Reduced dimension

Instance Members

▸ fit(x)

Fit model.

fit(x: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

▸ predict(x)

Returns reduced datas.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

AnomalyPCA

Principal component analysis for anomaly detection

new AnomalyPCA()

Extends PCA

Instance Members

▸ fit(x)

Fit model.

fit(x: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

▸ predict(x)

Returns anomaly degrees.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

PossibilisticCMeans

Possibilistic c-means

new PossibilisticCMeans(m: number?)

Parameters

m (number? = 2) Fuzziness factor

Instance Members

▸ init(datas)

Initialize model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ add()

Add a new cluster.

add()

▸ fit()

Fit model.

fit()

▸ predict()

Returns predicted coefficients.

predict(): Array<Array<number>>

Returns

Array<Array<number>>: Predicted values

PCR

lib/model/pcr.js

Principal component regression

new PCR()

Instance Members

▸ fit(x, y)

lib/model/pcr.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

lib/model/pcr.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

Pegasos

Primal Estimated sub-GrAdientSOlver for SVM

new Pegasos(rate: number, k: number?)

Parameters

rate (number) Learning rate

k (number? = 1) Batch size

Instance Members

▸ init(train_x, train_y)

Initialize this model.

init(train_x: Array<Array<number>>, train_y: Array<(1 | -1)>)

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<(1 | -1)>) Target values

▸ update(x, y)

Update model parameters with some data.

update(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target value

▸ fit()

Fit model parameters.

fit()

▸ predict(data)

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

PercentileAnormaly

lib/model/percentile.js

Percentile anomaly detection

new PercentileAnormaly(percentile: number, distribution: ("data" | "normal")?)

Parameters

percentile (number) Percentile value

distribution (("data" | "normal")? = 'data') Distribution name

Instance Members

▸ fit(data)

lib/model/percentile.js

Fit model.

fit(data: Array<Array<number>>)

Parameters

data (Array<Array<number>>) Training data

▸ predict(x)

lib/model/percentile.js

Returns predicted anomaly flags.

predict(x: Array<Array<number>>): Array<boolean>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<boolean>: true if a data is anomaly.

Perceptron

Perceptron

new Perceptron(rate: number)

Parameters

rate (number) Learning rate

Instance Members

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

AveragedPerceptron

Averaged perceptron

new AveragedPerceptron(rate: number)

Parameters

rate (number) Learning rate

Instance Members

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

MulticlassPerceptron

Multiclass perceptron

new MulticlassPerceptron(rate: number)

Parameters

rate (number) Learning rate

Instance Members

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<any>)

Parameters

x (Array<Array<number>>) Training data

y (Array<any>) Target values

▸ predict(data)

Returns predicted values.

predict(data: Array<Array<number>>): Array<any>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

PhansalkarThresholding

lib/model/phansalkar.js

Phansalkar thresholding

new PhansalkarThresholding(n: number?, k: number?, r: number?, p: number?, q: number?)

Parameters

n (number? = 3) Size of local range

k (number? = 0.25) Tuning parameter

r (number? = 0.5) Tuning parameter

p (number? = 2) Tuning parameter

q (number? = 10) Tuning parameter

Instance Members

▸ predict(x)

lib/model/phansalkar.js

Returns thresholded values.

predict(x: Array<Array<number>>): Array<Array<(0 | 1)>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<(0 | 1)>>: Predicted values

PLS

Partial least squares regression

new PLS(l: number)

Parameters

l (number) Limit on the number of latent factors

Instance Members

▸ init(x, y)

Initialize model.

init(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ fit()

Fit model.

fit()

▸ predict(x)

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

PLSA

Probabilistic latent semantic analysis

new PLSA(k: number?)

Parameters

k (number? = 2) Number of clusters

Instance Members

▸ init(x)

Initialize model.

init(x: Array<Array<any>>)

Parameters

x (Array<Array<any>>) Training data

▸ fit()

Fit model.

fit()

▸ predict()

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

PoissonRegression

lib/model/poisson.js

Poisson regression

new PoissonRegression(rate: number)

Parameters

rate (number) Learning rate

Instance Members

▸ fit(x, y)

lib/model/poisson.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

lib/model/poisson.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

PGAgent

Policy gradient agent

new PGAgent(env: RLEnvironmentBase, resolution: number?)

Parameters

env (RLEnvironmentBase) Environment

resolution (number? = 20) Resolution

Instance Members

▸ reset()

Reset agent.

reset()

▸ get_score()

Returns a score.

get_score(): Array<Array<Array<number>>>

Returns

Array<Array<Array<number>>>: Score values

▸ get_action(state)

Returns a action.

get_action(state: Array<any>): Array<any>

Parameters

state (Array<any>) Current states

Returns

Array<any>: Action

▸ update(action, state, reward, done, learning_rate)

lib/model/polynomial_histogram.js

Update model.

update(action: Array<any>, state: Array<any>, reward: number, done: boolean, learning_rate: number)

Parameters

action (Array<any>) Action

state (Array<any>) Next states

reward (number) Reward

done (boolean) Done epoch or not

learning_rate (number) Learning rate

PolynomialHistogram

Polynomial histogram

new PolynomialHistogram(p: number?, h: number?)

Parameters

p (number? = 2) Order

h (number? = 0.1) Bin size

Instance Members

▸ fit(x)

lib/model/polynomial_histogram.js

Fit model.

fit(x: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

▸ predict(x)

lib/model/polynomial_histogram.js

Returns predicted dencity.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

PolynomialInterpolation

lib/model/polynomial_interpolation.js

Polynomial interpolation

new PolynomialInterpolation()

Instance Members

▸ fit(x, y)

lib/model/polynomial_interpolation.js

Fit model.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(x)

lib/model/polynomial_interpolation.js

Returns predicted interpolated values.

predict(x: Array<number>): Array<number>

Parameters

x (Array<number>) Sample data

Returns

Array<number>: Predicted values

ProjectionPursuit

lib/model/ppr.js

Projection pursuit regression

new ProjectionPursuit(r: number?)

Parameters

r (number? = 5) Number of functions

Instance Members

▸ fit(x, y)

lib/model/ppr.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

lib/model/ppr.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

PRank

lib/model/prank.js

Perceptron ranking

new PRank(rate: number?)

Parameters

rate (number? = 0.1) Learning rate

Instance Members

▸ fit(x, y)

lib/model/prank.js

Fit model.

fit(x: Array<Array<number>>, y: Array<number>)

Parameters

x (Array<Array<number>>) Training data

y (Array<number>) Target values

▸ predict(x)

lib/model/prank.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

Prewitt

lib/model/prewitt.js

Prewitt edge detection

new Prewitt(th: number)

Parameters

th (number) Threshold

Instance Members

▸ predict(x)

lib/model/prewitt.js

Returns predicted edge flags.

predict(x: Array<Array<number>>): Array<Array<boolean>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<boolean>>: Predicted values. true if a pixel is edge.

PriestleyChao

lib/model/priestley_chao.js

Priestley–Chao kernel estimator

new PriestleyChao(h: number)

Parameters

h (number) Smoothing parameter for the kernel

Instance Members

▸ fit(x, y)

lib/model/priestley_chao.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

lib/model/priestley_chao.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

PrincipalCurve

lib/model/principal_curve.js

Principal curves

new PrincipalCurve()

Instance Members

▸ fit(x)

lib/model/principal_curve.js

Fit model.

fit(x: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

▸ predict()

lib/model/principal_curve.js

Returns reduced datas.

predict(): Array<Array<number>>

Returns

Array<Array<number>>: Predicted values

ProbabilisticPCA

lib/model/probabilistic_pca.js

Probabilistic Principal component analysis

new ProbabilisticPCA(method: ("analysis" | "em" | "bayes")?, rd: number)

Parameters

method (("analysis" | "em" | "bayes")? = 'analysis') Method name

rd (number) Reduced dimension

Instance Members

▸ fit(x)

lib/model/probabilistic_pca.js

Fit model.

fit(x: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

▸ predict(x)

lib/model/probabilistic_pca.js

Returns reduced datas.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

ProbabilityModel

ProbabilityModel

Type: object

Properties

fit (function (Array<Array<number>>): void) : Fit model

probability (function (Array<Array<number>>): Array<number>) : Returns predicted values

ProbabilityBasedClassifier

Probability based classifier

new ProbabilityBasedClassifier(model: any)

Parameters

model (any)

Instance Members

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<any>)

Parameters

x (Array<Array<number>>) Training data

y (Array<any>) Target values

▸ predict(x)

Returns predicted values.

predict(x: Array<Array<number>>): Array<(any | null)>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<(any | null)>: Predicted values

Probit

Probit

new Probit()

Instance Members

▸ init(train_x, train_y)

Initialize this model.

init(train_x: Array<Array<number>>, train_y: Array<(1 | -1)>)

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<(1 | -1)>) Target values

▸ fit()

Fit model parameters.

fit()

▸ predict(data)

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

MultinomialProbit

Multinomial probit

new MultinomialProbit()

Extends Probit

Instance Members

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<any>)

Parameters

x (Array<Array<number>>) Training data

y (Array<any>) Target values

▸ predict(data)

Returns predicted categories.

predict(data: Array<Array<number>>): Array<any>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

PROCLUS

PROjected CLUStering algorithm

new PROCLUS(k: number, a: number, b: number, l: number, minDeviation: number?)

Parameters

k (number) Number of clusters

a (number) Number to multiply the number of clusters for sample size

b (number) Number to multiply the number of clusters for final set size

l (number) Average dimensions

minDeviation (number? = 0.1) Minimum deviation to check the medoid is bad

Instance Members

▸ init(datas)

Initialize model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ fit()

Fit model.

fit()

▸ predict()

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

▸ outliers()

Returns a list of the data predicted as outliers or not.

outliers(): Array<boolean>

Returns

Array<boolean>: Predicted values

Projectron

Projectron

new Projectron(eta: number?, kernel: any)

Parameters

eta (number? = 0) Threshold

kernel (any = 'gaussian')

Instance Members

▸ fit(x, y)

Fit model parameters.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

Projectronpp

Projectron++

new Projectronpp(eta: number?, kernel: any)

Parameters

eta (number? = 0) Threshold

kernel (any = 'gaussian')

Instance Members

▸ fit(x, y)

Fit model parameters.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

PTile

lib/model/ptile.js

P-tile thresholding

new PTile(p: number?)

Parameters

p (number? = 0.5) Percentile value

Instance Members

▸ predict(x)

lib/model/ptile.js

Returns thresholded values.

predict(x: Array<number>): Array<(0 | 1)>

Parameters

x (Array<number>) Training data

Returns

Array<(0 | 1)>: Predicted values

QTableBase

Base class for Q-table

new QTableBase(env: RLEnvironmentBase, resolution: number?)

Parameters

env (RLEnvironmentBase) Environment

resolution (number? = 20) Resolution

Instance Members

▸ tensor

Tensor

tensor

Type: Tensor

▸ states

States

states

Type: Array<(Array<any> | RLRealRange | RLIntRange)>

▸ actions

Actions

actions

Type: Array<(Array<any> | RLRealRange | RLIntRange)>

▸ resolution

Resolution

resolution

Type: number

▸ toArray()

Returns Q-table as array.

toArray(): Array<any>

Returns

Array<any>: Nested array

▸ best_action(state)

Returns the best action.

best_action(state: Array<any>): Array<any>

Parameters

state (Array<any>) Current states

Returns

Array<any>: Action

QAgent

Q-learning agent

new QAgent(env: RLEnvironmentBase, resolution: number?)

Parameters

env (RLEnvironmentBase) Environment

resolution (number? = 20) Resolution

Instance Members

▸ get_score()

Returns a score.

get_score(): Array<Array<Array<number>>>

Returns

Array<Array<Array<number>>>: Score values

▸ get_action(state, greedy_rate)

Returns a action.

get_action(state: Array<any>, greedy_rate: number): Array<any>

Parameters

state (Array<any>) Current states

greedy_rate (number = 0.002) Greedy rate

Returns

Array<any>: Action

▸ update(action, state, next_state, reward)

lib/model/quadratic_discriminant.js

Update model.

update(action: Array<any>, state: Array<any>, next_state: Array<any>, reward: number)

Parameters

action (Array<any>) Action

state (Array<any>) Current state

next_state (Array<any>) Next state

reward (number) Reward

QuadraticDiscriminant

Quadratic discriminant analysis

new QuadraticDiscriminant()

Instance Members

▸ fit(x, y)

lib/model/quadratic_discriminant.js

Fit model.

fit(x: Array<Array<number>>, y: Array<any>)

Parameters

x (Array<Array<number>>) Training data

y (Array<any>) Target values

▸ predict(data)

lib/model/quadratic_discriminant.js

Returns predicted categories.

predict(data: Array<Array<number>>): Array<any>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

QuantileRegression

lib/model/quantile_regression.js

Quantile regression

new QuantileRegression(tau: number?)

Parameters

tau (number? = 0.5) Quantile value

Instance Members

▸ fit(x, y, learningRate = 0.1)

lib/model/quantile_regression.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>, learningRate: number?)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

learningRate (number? = 0.1) Learning rate

▸ predict(x)

lib/model/quantile_regression.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

RadiusNeighborBase

Bsae class for radius neighbor models

new RadiusNeighborBase(r: number?, metric: ("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)?)

Parameters

r (number? = 1) Radius to determine neighborhood

metric (("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)? = 'euclid') Metric name

Instance Members

▸ _add(point, category?)

Add a data.

_add(point: Array<number>, category: any?)

Parameters

point (Array<number>) Training data

category (any?) Target value

RadiusNeighbor

radius neighbor

new RadiusNeighbor(r: number?, metric: ("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)?)

Extends RadiusNeighborBase

Parameters

r (number? = 1) Radius to determine neighborhood

metric (("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)? = 'euclid') Metric name

Instance Members

▸ add(point, category)

Add a data.

add(point: Array<number>, category: any)

Parameters

point (Array<number>) Training data

category (any) Target value

▸ fit(datas, targets)

Add datas.

fit(datas: Array<Array<number>>, targets: Array<any>)

Parameters

datas (Array<Array<number>>) Training data

targets (Array<any>) Target values

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<any>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

RadiusNeighborRegression

radius neighbor regression

new RadiusNeighborRegression(r: number?, metric: ("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)?)

Extends RadiusNeighborBase

Parameters

r (number? = 1) Radius to determine neighborhood

metric (("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)? = 'euclid') Metric name

Instance Members

▸ add(point, category)

Add a data.

add(point: Array<number>, category: number)

Parameters

point (Array<number>) Training data

category (number) Target value

▸ fit(datas, targets)

Add datas.

fit(datas: Array<Array<number>>, targets: Array<number>)

Parameters

datas (Array<Array<number>>) Training data

targets (Array<number>) Target values

▸ predict(datas)

Returns predicted values.

predict(datas: Array<Array<number>>): Array<(number | null)>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<(number | null)>: Predicted values

SemiSupervisedRadiusNeighbor

Semi-supervised radius neighbor

new SemiSupervisedRadiusNeighbor(k: number?, metric: ("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)?)

Extends RadiusNeighborBase

Parameters

k (number? = 5) Radius to determine neighborhood

metric (("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)? = 'euclid') Metric name

Instance Members

▸ add(point, category)

Add a data.

add(point: Array<number>, category: (any | null))

Parameters

point (Array<number>) Training data

category ((any | null)) Target value

▸ fit(datas, targets)

Add datas.

fit(datas: Array<Array<number>>, targets: Array<(any | null)>)

Parameters

datas (Array<Array<number>>) Training data

targets (Array<(any | null)>) Target values

▸ predict()

lib/model/ramer_douglas_peucker.js

Returns predicted values.

predict(): Array<any>

Returns

Array<any>: Predicted values

RamerDouglasPeucker

Ramer-Douglas-Peucker algorithm

new RamerDouglasPeucker(e: number?)

Parameters

e (number? = 0.1) Threshold of distance

Instance Members

▸ fit(x, y)

lib/model/ramer_douglas_peucker.js

Fit model.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(x)

lib/model/ramer_douglas_peucker.js

Returns predicted values.

predict(x: Array<number>): Array<number>

Parameters

x (Array<number>) Sample data

Returns

Array<number>: Predicted values

RandomForest

Bsae class for random forest models

new RandomForest(tree_num: number, sampling_rate: number?, tree_class: (DecisionTreeClassifier | DecisionTreeRegression), tree_class_args: Array<any>?)

Parameters

tree_num (number) Number of trees

sampling_rate (number? = 0.8) Sampling rate

tree_class ((DecisionTreeClassifier | DecisionTreeRegression)) Tree class

tree_class_args (Array<any>? = null) Arguments for constructor of tree class

Instance Members

▸ depth

The max depth among the trees.

depth

Type: number

▸ init(datas, targets)

init(datas: Array<Array<number>>, targets: Array<any>)

Parameters

datas (Array<Array<number>>) Training data

targets (Array<any>) Target values

▸ fit()

Fit model.

fit()

▸ predict_prob(datas)

Returns probability of the datas.

predict_prob(datas: Array<Array<number>>): Array<Map<number, number>>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<Map<number, number>>: Predicted values

RandomForestClassifier

Random forest classifier

new RandomForestClassifier(tree_num: number, sampling_rate: number?, method: ("ID3" | "CART")?)

Extends RandomForest

Parameters

tree_num (number) Number of trees

sampling_rate (number? = 0.8) Sampling rate

method (("ID3" | "CART")? = 'CART') Method name

Instance Members

▸ predict(datas)

Returns predicted values.

predict(datas: Array<Array<number>>): Array<any>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

RandomForestRegressor

Random forest regressor

new RandomForestRegressor(tree_num: number, sampling_rate: number?)

Extends RandomForest

Parameters

tree_num (number) Number of trees

sampling_rate (number? = 0.8) Sampling rate

Instance Members

▸ predict(datas)

lib/model/random_projection.js

Returns predicted values.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

RandomProjection

Random projection

new RandomProjection(init: ("uniform" | "root3" | "normal")?, rd: (number | null)?)

Parameters

init (("uniform" | "root3" | "normal")? = 'uniform') Initialize method name

rd ((number | null)? = null) Reduced dimension

Instance Members

▸ predict(x)

lib/model/random_projection.js

Returns reduced values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<number>>: Predicted values

RankNet

lib/model/ranknet.js

RankNet

new RankNet(layer_sizes: Array<number>, activations: (string | Array<string>)?, rate: number?)

Parameters

layer_sizes (Array<number>) Sizes of all layers

activations ((string | Array<string>)? = 'tanh') Activation names

rate (number? = 0.01) Learning rate

Instance Members

▸ fit(x1, x2, comp)

lib/model/ranknet.js

Fit model.

fit(x1: Array<Array<number>>, x2: Array<Array<number>>, comp: Array<(-1 | 0 | 1)>): number

Parameters

x1 (Array<Array<number>>) Training data 1

x2 (Array<Array<number>>) Training data 2

comp (Array<(-1 | 0 | 1)>) Sign of (data 1 rank - data 2 rank). If data 1 rank is bigger than data 2, corresponding value is 1.

Returns

number: loss

▸ predict(x)

lib/model/ranknet.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

RANSACSubModel

RANSACSubModel

Type: object

Properties

fit (function (Array<Array<number>>, Array<any>): void) : Fit model

predict (function (Array<Array<number>>): Array<any>) : Returns predicted values

score (function (Array<any>, Array<any>): number?) : Returns a number how accurate the prediction is

RANSAC

Random sample consensus

new RANSAC(model: any, sample: (number | null)?)

Parameters

model (any)

sample ((number | null)? = null) Sampling rate

Instance Members

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<any>)

Parameters

x (Array<Array<number>>) Training data

y (Array<any>) Target values

▸ predict(x)

Returns predicted values.

predict(x: Array<Array<number>>): Array<any>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

RadialBasisFunctionNetwork

lib/model/rbf.js

Radial basis function network

Parameters

rbf

(("linear" | "gaussian" | "multiquadric" | "inverse quadratic" | "inverse multiquadric" | "thin plate" | "bump")? = 'linear')

RBF name

e (number? = 1) Tuning parameter

l (number? = 0) Regularization parameter

Instance Members

▸ fit(x, y)

lib/model/rbf.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(target)

lib/model/rbf.js

Returns predicted values.

predict(target: Array<Array<number>>): Array<number>

Parameters

target (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

RBM

Restricted Boltzmann machine

new RBM(hiddenSize: number, lr: number?)

Parameters

hiddenSize (number) Size of hidden layer

lr (number? = 0.01) Learning rate

Instance Members

▸ fit(x)

Fit model.

fit(x: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

▸ energy(v, h)

Return a energy value of the data.

energy(v: Array<number>, h: Array<number>): number

Parameters

v (Array<number>) Sample data

h (Array<number>) Target values

Returns

number: Energy value

▸ predict(x)

Returns predicted values.

predict(x: Array<number>): Array<(0 | 1)>

Parameters

x (Array<number>) Sample data

Returns

Array<(0 | 1)>: Predicted values

GBRBM

Gaussian-Bernouili Restricted Boltzmann machine

new GBRBM(hiddenSize: number, lr: number?, fixSigma: boolean?)

Parameters

hiddenSize (number) Size of hidden layer

lr (number? = 0.01) Learning rate

fixSigma (boolean? = false) Do not learn sigma or not

Instance Members

▸ fit(x)

Fit model.

fit(x: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

▸ energy(v, h)

Return a energy value of the data.

energy(v: Array<number>, h: Array<number>): number

Parameters

v (Array<number>) Sample data

h (Array<number>) Target values

Returns

number: Energy value

▸ predict(x)

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

RBP

lib/model/rbp.js

Randomized Budget Perceptron

new RBP(b: number)

Parameters

b (number) Number of support vectors

Instance Members

▸ fit(x, y)

lib/model/rbp.js

Fit model parameters.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

lib/model/rbp.js

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

RDF

lib/model/rdf.js

Relative Density Factor

new RDF(r: number?)

Parameters

r (number? = 1.0) Radius

Instance Members

▸ predict(datas)

lib/model/rdf.js

Returns anomaly degrees.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Predicted values

RDOS

lib/model/rdos.js

Relative Density-based Outlier Score

new RDOS(k: number, h: number, kernel: ("gaussian" | {name: "gaussian"} | function (Array<number>): number)?)

Parameters

k (number) Number of neighborhoods

h (number) Kernel width

kernel (("gaussian" | {name: "gaussian"} | function (Array<number>): number)? = 'gaussian') Kernel name

Instance Members

▸ predict(datas)

lib/model/rdos.js

Returns anomaly degrees.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Predicted values

Ridge

Ridge regressioin

new Ridge(lambda: number?)

Parameters

lambda (number? = 0.1) Regularization strength

Instance Members

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

▸ importance()

Returns importances of the features.

importance(): Array<number>

Returns

Array<number>: Importances

MulticlassRidge

Multiclass ridge regressioin

new MulticlassRidge(lambda: number?)

Parameters

lambda (number? = 0.1) Regularization strength

Instance Members

▸ categories

Category list

categories

Type: Array<any>

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<any>)

Parameters

x (Array<Array<number>>) Training data

y (Array<any>) Target values

▸ predict(x)

Returns predicted values.

predict(x: Array<Array<number>>): Array<any>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

▸ importance()

Returns importances of the features.

importance(): Array<number>

Returns

Array<number>: Importances

KernelRidge

Kernel ridge regression

new KernelRidge(lambda: number?, kernel: any)

Parameters

lambda (number? = 0.1) Regularization strength

kernel (any = 'gaussian')

Instance Members

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

▸ importance()

Returns importances of the features.

importance(): Array<number>

Returns

Array<number>: Importances

RKOF

lib/model/rkof.js

Robust Kernel-based Outlier Factor

new RKOF(k: number, h: number, alpha: number, kernel: any)

Parameters

k (number) Number of neighborhoods

h (number) Smoothing parameter

alpha (number) Sensitivity parameter

kernel (any = 'gaussian')

Instance Members

▸ predict(datas)

lib/model/rkof.js

Returns anomaly degrees.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Predicted values

RecursiveLeastSquares

Recursive least squares

new RecursiveLeastSquares()

Instance Members

▸ update(x, y)

Update model parameters with one data.

update(x: Matrix, y: number)

Parameters

x (Matrix) Training data

y (number) Target value

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<number>)

Parameters

x (Array<Array<number>>) Training data

y (Array<number>) Target values

▸ predict(data)

Returns predicted datas.

predict(data: Array<Array<number>>): Array<number>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

RepeatedMedianRegression

lib/model/rmr.js

Repeated median regression

new RepeatedMedianRegression()

Instance Members

▸ fit(x, y)

lib/model/rmr.js

Fit model.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(x)

lib/model/rmr.js

Returns predicted values.

predict(x: Array<number>): Array<number>

Parameters

x (Array<number>) Sample data

Returns

Array<number>: Predicted values

RNN

Recurrent neuralnetwork

new RNN(method: ("rnn" | "lstm" | "gru")?, window: number?, unit: number?, out_size: number?, optimizer: string?)

Parameters

method (("rnn" | "lstm" | "gru")? = 'lstm') Method name

window (number? = 10) Window size

unit (number? = 10) Size of recurrent unit

out_size (number? = 1) Output size

optimizer (string? = 'adam') Optimizer of the network

Instance Members

▸ method

Method

method

Type: ("rnn" | "lstm" | "gru")

▸ epoch

Epoch

epoch

Type: number

▸ fit(train_x, train_y, iteration, rate, batch)

Fit model.

fit(train_x: Array<Array<number>>, train_y: Array<Array<number>>, iteration: number, rate: number, batch: number): number

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<Array<number>>) Target values

iteration (number) Iteration count

rate (number) Learning rate

batch (number) Batch size

Returns

number: Loss value

▸ predict(data, k)

lib/model/robust_scaler.js

Returns predicted future values.

predict(data: Array<Array<number>>, k: number): Array<Array<number>>

Parameters

data (Array<Array<number>>) Sample data

k (number) Prediction count

Returns

Array<Array<number>>: Predicted values

RobertsCross

lib/model/roberts.js

Roberts cross

new RobertsCross(th: number)

Parameters

th (number) Threshold

Instance Members

▸ predict(x)

lib/model/roberts.js

Returns predicted edge flags.

predict(x: Array<Array<number>>): Array<Array<boolean>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<boolean>>: true if a pixel is edge.

RobustScaler

Robust scaler

new RobustScaler()

Instance Members

▸ fit(x)

lib/model/robust_scaler.js

Fit model.

fit(x: (Array<number> | Array<Array<number>>))

Parameters

x ((Array<number> | Array<Array<number>>)) Training data

▸ predict(x)

lib/model/robust_scaler.js

Returns transformed values.

predict(x: (Array<number> | Array<Array<number>>)): (Array<number> | Array<Array<number>>)

Parameters

x ((Array<number> | Array<Array<number>>)) Sample data

Returns

(Array<number> | Array<Array<number>>): Predicted values

ROCKNode

ROCKNode

Type: object

Properties

point (Array<number>?) : Data point of leaf node

index (number?) : Data index of leaf node

g (number) : Number of leaf nodes

distance (number?) : Distance between children nodes

children (Array<ROCKNode>?) : Children nodes

leafs (Array<ROCKNode>) : Leaf nodes

ROCK

RObust Clustering using linKs

new ROCK(th: number, k: number)

Parameters

th (number) Threshold

k (number) Number of clusters

Instance Members

▸ fit(data)

Fit model.

fit(data: Array<Array<number>>)

Parameters

data (Array<Array<number>>) Training data

▸ getClusters(number)

Returns the specified number of clusters.

getClusters(number: number): Array<ROCKNode>

Parameters

number (number) Number of clusters

Returns

Array<ROCKNode>: Cluster nodes

▸ predict()

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

ROMMA

Relaxed Online Maximum Margin Algorithm

new ROMMA()

Instance Members

▸ update(x, y)

Update model parameters with one data.

update(x: Array<number>, y: (1 | -1))

Parameters

x (Array<number>) Training data

y ((1 | -1)) Target value

▸ fit(x, y)

Fit model parameters.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

Returns predicted datas.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

AggressiveROMMA

Aggressive Relaxed Online Maximum Margin Algorithm

new AggressiveROMMA()

Extends ROMMA

RVM

lib/model/rvm.js

Relevance vector machine

new RVM()

Instance Members

▸ fit(x, y)

lib/model/rvm.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

lib/model/rvm.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

S3VM

Semi-Supervised Support Vector Machine

new S3VM(kernel: any)

Parameters

kernel (any)

Instance Members

▸ init(x, y)

Initialize model.

init(x: Array<Array<number>>, y: Array<(1 | -1 | null)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1 | null)>) Target values

▸ fit()

Fit model.

fit()

▸ predict(data)

Returns predicted values.

predict(data: Array<Array<number>>): Array<number>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

Sammon

lib/model/sammon.js

Sammon mapping

new Sammon(x: Array<Array<number>>, rd: number)

Parameters

x (Array<Array<number>>) Sample data

rd (number) Reduced dimension

Instance Members

▸ fit()

lib/model/sammon.js

Fit model and returns reduced values.

fit(): Array<Array<number>>

Returns

Array<Array<number>>: Predicted values

▸ predict()

lib/model/sammon.js

Returns reduced values.

predict(): Array<Array<number>>

Returns

Array<Array<number>>: Predicted values

SARSAAgent

SARSA agent

new SARSAAgent(env: RLEnvironmentBase, resolution: number?)

Parameters

env (RLEnvironmentBase) Environment

resolution (number? = 20) Resolution

Instance Members

▸ reset()

Reset agent.

reset()

▸ get_score()

Returns a score.

get_score(): Array<Array<Array<number>>>

Returns

Array<Array<Array<number>>>: Score values

▸ get_action(state, greedy_rate)

Returns a action.

get_action(state: Array<any>, greedy_rate: number): Array<any>

Parameters

state (Array<any>) Current states

greedy_rate (number = 0.002) Greedy rate

Returns

Array<any>: Action

▸ update(action, state, next_state, reward)

lib/model/savitzky_golay.js

Update model.

update(action: Array<any>, state: Array<any>, next_state: Array<any>, reward: number)

Parameters

action (Array<any>) Action

state (Array<any>) Current states

next_state (Array<any>) Next states

reward (number) Reward

SauvolaThresholding

lib/model/sauvola.js

sauvola thresholding

new SauvolaThresholding(n: number?, k: number?, r: number?)

Parameters

n (number? = 3) Size of local range

k (number? = 0.1) Tuning parameter

r (number? = 1) Tuning parameter

Instance Members

▸ predict(x)

lib/model/sauvola.js

Returns thresholded values.

predict(x: Array<Array<number>>): Array<Array<(0 | 1)>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<(0 | 1)>>: Predicted values

SavitzkyGolayFilter

Savitzky-Golay filter

new SavitzkyGolayFilter(k: number)

Parameters

k (number) Number of coefficients

Instance Members

▸ predict(x)

lib/model/savitzky_golay.js

Returns predicted values.

predict(x: Array<number>): Array<number>

Parameters

x (Array<number>) Training data

Returns

Array<number>: Predicted values

SDAR

lib/model/sdar.js

Sequentially Discounting Autoregressive model

new SDAR(p: number?, r: number?)

Deprecated: Does not work properly

Parameters

p (number? = 1) Order

r (number? = 0.8) Forgetting factor

Instance Members

▸ probability(data)

lib/model/sdar.js

Returns probability of the datas.

probability(data: Array<number>): Array<number>

Parameters

data (Array<number>) Sample data

Returns

Array<number>: Predicted values

▸ predict(data, k)

lib/model/sdar.js

Returns predicted future values.

predict(data: Array<number>, k: number): Array<number>

Parameters

data (Array<number>) Training data

k (number) Prediction count

Returns

Array<number>: Predicted values

SegmentedRegression

lib/model/segmented.js

Segmented regression

new SegmentedRegression(seg: number?)

Parameters

seg (number? = 3) Number of segments

Instance Members

▸ fit(x, y)

lib/model/segmented.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

lib/model/segmented.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

SelectiveNaiveBayes

lib/model/selective_naive_bayes.js

Selective Naive bayes

new SelectiveNaiveBayes(distribution: "gaussian"?)

Parameters

distribution ("gaussian"? = 'gaussian') Distribution name

Instance Members

▸ fit(datas, labels)

lib/model/selective_naive_bayes.js

Fit model.

fit(datas: Array<Array<number>>, labels: Array<any>)

Parameters

datas (Array<Array<number>>) Training data

labels (Array<any>) Target values

▸ predict(data)

lib/model/selective_naive_bayes.js

Returns predicted categories.

predict(data: Array<Array<number>>): Array<any>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

SelectiveSamplingPerceptron

Selective sampling Perceptron

new SelectiveSamplingPerceptron(b: number, rate: number)

Parameters

b (number) Smooth parameter

rate (number) Learning rate

Instance Members

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

SelectiveSamplingAdaptivePerceptron

Selective sampling Perceptron with adaptive parameter

new SelectiveSamplingAdaptivePerceptron(beta: number, rate: number)

Parameters

beta (number) Smooth parameter

rate (number) Learning rate

Instance Members

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

SelectiveSamplingSOP

Selective sampling second-order Perceptron

new SelectiveSamplingSOP(b: number)

Parameters

b (number) Smooth parameter

Instance Members

▸ init(train_x, train_y)

Initialize this model.

init(train_x: Array<Array<number>>, train_y: Array<(1 | -1)>)

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<(1 | -1)>) Target values

▸ update(x, y)

Update model parameters with one data.

update(x: Matrix, y: (1 | -1))

Parameters

x (Matrix) Training data

y ((1 | -1)) Target value

▸ fit()

Fit model parameters.

fit()

▸ predict(data)

lib/model/selective_sampling_winnow.js

Returns predicted datas.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

SelectiveSamplingWinnow

Selective sampling Winnow

new SelectiveSamplingWinnow(b: number, alpha: boolean?)

Parameters

b (number) Smooth parameter

alpha (boolean? = 2) Learning rate

Instance Members

▸ fit(x, y)

lib/model/selective_sampling_winnow.js

Fit model.

fit(x: Array<Array<(1 | -1)>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<(1 | -1)>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

lib/model/selective_sampling_winnow.js

Returns predicted values.

predict(data: Array<Array<(1 | -1)>>): Array<(1 | -1)>

Parameters

data (Array<Array<(1 | -1)>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

SelfTraining

Self-training

new SelfTraining(model: object, threshold: number)

Parameters

model (object) View

threshold (number) Threshold

Instance Members

▸ init(x, y)

Initialize model.

init(x: Array<Array<number>>, y: Array<(any | null)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(any | null)>) Target values

▸ fit()

Fit model.

fit()

▸ predict()

Returns predicted categories.

predict(): Array<(any | null)>

Returns

Array<(any | null)>: Predicted values

SemiSupervisedNaiveBayes

Semi-supervised naive bayes

new SemiSupervisedNaiveBayes(lambda: number?)

Parameters

lambda (number? = 1) Weight applied to the contribution of the unlabeled data

Instance Members

▸ init(datas, labels)

Initialize model.

init(datas: Array<Array<string>>, labels: Array<(any | null)>)

Parameters

datas (Array<Array<string>>) Training data

labels (Array<(any | null)>) Target values

▸ fit()

Fit model.

fit()

▸ probability(datas)

Returns predicted probabilities.

probability(datas: Array<Array<string>>): Array<Array<number>>

Parameters

datas (Array<Array<string>>) Sample data

Returns

Array<Array<number>>: Predicted values

▸ logLikelihood()

Returns predicted categories.

logLikelihood(): number

Returns

number: Log likelihood value

▸ predict(datas)

lib/model/shifting_perceptron.js

Returns predicted categories.

predict(datas: Array<Array<string>>): Array<any>

Parameters

datas (Array<Array<string>>) Sample data

Returns

Array<any>: Predicted values

SezanThresholding

lib/model/sezan.js

Sezan's thresholding

new SezanThresholding(gamma: number?, sigma: number?)

Parameters

gamma (number? = 0.5) Tradeoff value between black and white

sigma (number? = 5) Sigma of normal distribution

Instance Members

▸ predict(x)

lib/model/sezan.js

Returns thresholded values.

predict(x: Array<number>): Array<(0 | 1)>

Parameters

x (Array<number>) Training data

Returns

Array<(0 | 1)>: Predicted values

ShiftingPerceptron

Shifting Perceptron Algorithm

new ShiftingPerceptron(lambda: number)

Parameters

lambda (number) Rate of weight decay

Instance Members

▸ fit(x, y)

lib/model/shifting_perceptron.js

Fit model parameters.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

lib/model/shifting_perceptron.js

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

ILK

Implicit online Learning with Kernels

new ILK(eta: number?, lambda: number?, c: number?, kernel: any, loss: ("square" | "hinge" | "logistic")?)

Parameters

eta (number? = 1) Learning rate

lambda (number? = 1) Regularization constant

c (number? = 1) Penalty imposed on point prediction violations.

kernel (any = 'gaussian')

loss (("square" | "hinge" | "logistic")? = 'hinge') Loss type name

Instance Members

▸ update(x, y)

Update model parameters with one data.

update(x: Array<number>, y: (1 | -1))

Parameters

x (Array<number>) Training data

y ((1 | -1)) Target value

▸ fit(x, y)

Fit model.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

SILK

Sparse Implicit online Learning with Kernels

new SILK(eta: number?, lambda: number?, c: number?, w: number?, kernel: any, loss: ("square" | "hinge" | "graph" | "logistic")?)

Extends ILK

Parameters

eta (number? = 1) Learning rate

lambda (number? = 1) Regularization constant

c (number? = 1) Penalty imposed on point prediction violations.

w (number? = 10) Buffer size

kernel (any = 'gaussian')

loss (("square" | "hinge" | "graph" | "logistic")? = 'hinge') Loss type name

Instance Members

▸ update(x, y)

lib/model/sinc_interpolation.js

Update model parameters with one data.

update(x: Array<number>, y: (1 | -1))

Parameters

x (Array<number>) Training data

y ((1 | -1)) Target value

SincInterpolation

Sinc interpolation

new SincInterpolation()

Instance Members

▸ fit(values)

lib/model/sinc_interpolation.js

Fit model parameters.

fit(values: Array<number>)

Parameters

values (Array<number>) Training data

▸ predict(index)

lib/model/sinc_interpolation.js

Returns predicted interpolated values.

predict(index: Array<number>): Array<number>

Parameters

index (Array<number>) Sample data

Returns

Array<number>: Predicted values

SlicedInverseRegression

lib/model/sir.js

Sliced inverse regression

new SlicedInverseRegression(s: number, rd: (number | null)?)

Parameters

s (number) Number of slices

rd ((number | null)? = null) Reduced dimension

Instance Members

▸ predict(x, y)

lib/model/sir.js

Returns reduced datas.

predict(x: Array<Array<number>>, y: Array<number>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Training data

y (Array<number>) Target values

Returns

Array<Array<number>>: Predicted values

Slerp

lib/model/slerp.js

Spherical linear interpolation

new Slerp(o: number?)

Parameters

o (number? = 1) Angle subtended by the arc

Instance Members

▸ fit(x, y)

lib/model/slerp.js

Fit model.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(target)

lib/model/slerp.js

Returns predicted interpolated values.

predict(target: Array<number>): Array<number>

Parameters

target (Array<number>) Sample data

Returns

Array<number>: Predicted values

SliceSampling

lib/model/slice_sampling.js

slice sampling

new SliceSampling(targetFunc: function (Array<number>): number, d: number, w: number?)

Parameters

targetFunc (function (Array<number>): number) Target distribution

d (number) Output size

w (number? = 1.0) Check width

Instance Members

▸ sample(n)

lib/model/slice_sampling.js

Returns sampled values.

sample(n: number): Array<Array<number>>

Parameters

n (number) Number of generated data

Returns

Array<Array<number>>: Generated values

SMARegression

lib/model/sma.js

Standardizes Major Axis regression

new SMARegression()

Instance Members

▸ fit(x, y)

lib/model/sma.js

Fit model.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(x)

lib/model/sma.js

Returns predicted values.

predict(x: Array<number>): Array<number>

Parameters

x (Array<number>) Sample data

Returns

Array<number>: Predicted values

SmirnovGrubbs

lib/model/smirnov_grubbs.js

SmirnovGrubbs test

new SmirnovGrubbs(alpha: number)

Parameters

alpha (number) Significance level

Instance Members

▸ predict(data)

lib/model/smirnov_grubbs.js

Returns a list of the data predicted as outliers or not.

predict(data: Array<Array<number>>): Array<boolean>

Parameters

data (Array<Array<number>>) Training data

Returns

Array<boolean>: Predicted values

SmoothstepInterpolation

lib/model/smoothstep.js

Smoothstep interpolation

new SmoothstepInterpolation(n: number?)

Parameters

n (number? = 1) Order

Instance Members

▸ fit(x, y)

lib/model/smoothstep.js

Fit model.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(target)

lib/model/smoothstep.js

Returns predicted interpolated values.

predict(target: Array<number>): Array<number>

Parameters

target (Array<number>) Sample data

Returns

Array<number>: Predicted values

Snakes

Snakes (active contour model)

new Snakes(alpha: number, beta: number, gamma: number, k: number?)

Parameters

alpha (number) Penalty for length

beta (number) Penalty for curvature

gamma (number) Penalty for conformity with image

k (number? = 100) Number of vertices

Instance Members

▸ init(x)

Initialize model.

init(x: Array<Array<Array<number>>>)

Parameters

x (Array<Array<Array<number>>>) Training data

▸ fit()

Fit model.

fit()

▸ predict()

Returns predicted edge flags.

predict(): Array<Array<boolean>>

Returns

Array<Array<boolean>>: Predicted values. true if a pixel is edge.

Sobel

lib/model/sobel.js

Sobel edge detection

new Sobel(th: number)

Parameters

th (number) Threshold

Instance Members

▸ predict(x)

lib/model/sobel.js

Returns predicted edge flags.

predict(x: Array<Array<number>>): Array<Array<boolean>>

Parameters

x (Array<Array<number>>) Training data

Returns

Array<Array<boolean>>: Predicted values. true if a pixel is edge.

SoftKMeans

Soft k-means

new SoftKMeans(beta: number?)

Parameters

beta (number? = 1) Tuning parameter

Instance Members

▸ init(datas)

Initialize model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ add()

Add a new cluster.

add()

▸ fit()

Fit model.

fit()

▸ predict()

Returns predicted responsibilities.

predict(): Array<Array<number>>

Returns

Array<Array<number>>: Predicted values

SOM

lib/model/som.js

Self-Organizing Map

new SOM(input_size: number, output_size: number, resolution: number?)

Parameters

input_size (number) Input size

output_size (number) Output size

resolution (number? = 20) Resolution of output

Instance Members

▸ fit(data)

lib/model/som.js

Fit model.

fit(data: Array<Array<number>>)

Parameters

data (Array<Array<number>>) Training data

▸ predict(x)

lib/model/som.js

Returns predicted categories.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

SecondOrderPerceptron

Second order perceptron

new SecondOrderPerceptron(a: number?)

Parameters

a (number? = 1) Tuning parameter

Instance Members

▸ init(train_x, train_y)

Initialize this model.

init(train_x: Array<Array<number>>, train_y: Array<(1 | -1)>)

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<(1 | -1)>) Target values

▸ update(x, y)

Update model parameters with one data.

update(x: Matrix, y: (1 | -1))

Parameters

x (Matrix) Training data

y ((1 | -1)) Target value

▸ fit()

Fit model parameters.

fit()

▸ predict(data)

Returns predicted datas.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

SpectralClustering

Spectral clustering

new SpectralClustering(affinity: ("rbf" | "knn")?, param: object?)

Parameters

affinity (("rbf" | "knn")? = 'rbf') Affinity type name

param (object? = {}) Config

Instance Members

▸ size

Number of clusters.

size

Type: number

▸ epoch

Epoch.

epoch

Type: number

▸ init(datas)

Initialize model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ add()

Add a new cluster.

add()

▸ clear()

Clear all clusters.

clear()

▸ predict()

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

▸ fit()

lib/model/spline_interpolation.js

Fit and returns total distance the centroid has moved.

fit(): number

Returns

number: Total distance the centroid has moved

SmoothingSpline

lib/model/spline.js

Spline smoothing

new SmoothingSpline(l: number)

Parameters

l (number) Smoothing parameter

Instance Members

▸ fit(x, y)

lib/model/spline.js

Fit model.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(data)

lib/model/spline.js

Returns predicted values.

predict(data: Array<number>): Array<number>

Parameters

data (Array<number>) Sample data

Returns

Array<number>: Predicted values

SplineInterpolation

Spline interpolation

new SplineInterpolation()

Instance Members

▸ fit(x, y)

lib/model/spline_interpolation.js

Fit model.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(target)

lib/model/spline_interpolation.js

Returns predicted interpolated values.

predict(target: Array<number>): Array<number>

Parameters

target (Array<number>) Sample data

Returns

Array<number>: Predicted values

SplitAndMerge

lib/model/split_merge.js

Split and merge segmentation

new SplitAndMerge(method: ("variance" | "uniformity")?, threshold: number?)

Parameters

method (("variance" | "uniformity")? = 'variance') Method name

threshold (number? = 0.1) Threshold

Instance Members

▸ predict(x)

lib/model/split_merge.js

Returns predicted segments.

predict(x: Array<Array<Array<number>>>): Array<Array<number>>

Parameters

x (Array<Array<Array<number>>>) Training data

Returns

Array<Array<number>>: Predicted values

SquaredLossMICPD

lib/model/squared_loss_mi.js

Squared-loss Mutual information change point detection

new SquaredLossMICPD(model: object, w: number, take: number?, lag: number?)

Parameters

model (object) Density ratio estimation model

w (number) Window size

take (number?) Take number

lag (number?) Lag

Instance Members

▸ predict(datas)

lib/model/squared_loss_mi.js

Returns anomaly degrees.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

SST

lib/model/sst.js

Singular-spectrum transformation

new SST(w: number, take: number?, lag: number?)

Parameters

w (number) Window size

take (number?) Take number

lag (number?) Lag

Instance Members

▸ predict(datas)

lib/model/sst.js

Returns anomaly degrees.

predict(datas: Array<number>): Array<number>

Parameters

datas (Array<number>) Training data

Returns

Array<number>: Predicted values

Standardization

lib/model/standardization.js

Standardization

new Standardization(ddof: number?)

Parameters

ddof (number? = 0) Delta Degrees of Freedom

Instance Members

▸ fit(x)

lib/model/standardization.js

Fit model.

fit(x: (Array<number> | Array<Array<number>>))

Parameters

x ((Array<number> | Array<Array<number>>)) Training data

▸ predict(x)

lib/model/standardization.js

Returns transformed values.

predict(x: (Array<number> | Array<Array<number>>)): (Array<number> | Array<Array<number>>)

Parameters

x ((Array<number> | Array<Array<number>>)) Sample data

Returns

(Array<number> | Array<Array<number>>): Predicted values

StatisticalRegionMerging

lib/model/statistical_region_merging.js

Statistical Region Merging

new StatisticalRegionMerging(t: number)

Parameters

t (number) Threshold

Instance Members

▸ predict(x)

lib/model/statistical_region_merging.js

Fit model.

predict(x: Array<Array<Array<number>>>): Array<Array<Array<number>>>

Parameters

x (Array<Array<Array<number>>>) Training data

Returns

Array<Array<Array<number>>>: Predicted values

STING

lib/model/sting.js

STatistical INformation Grid-based method

new STING()

Deprecated: Not implemented

Instance Members

▸ fit(datas)

lib/model/sting.js

Fit model.

fit(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ predict(datas)

lib/model/sting.js

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

Stoptron

lib/model/stoptron.js

Stoptron

new Stoptron(n: number?, kernel: any)

Parameters

n (number? = 10) Cachs size

kernel (any = 'gaussian')

Instance Members

▸ fit(x, y)

lib/model/stoptron.js

Fit model.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

lib/model/stoptron.js

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

SVC

Support vector clustering

new SVC(kernel: any)

Parameters

kernel (any)

Instance Members

▸ size

Number of clusters

size

Type: number

▸ init(x)

Initialize this model.

init(x: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

▸ fit()

Fit model.

fit()

▸ predict()

Returns predicted categories.

predict(): Array<number>

Returns

Array<number>: Predicted values

SVM

Support vector machine

new SVM(kernel: any)

Parameters

kernel (any)

Instance Members

▸ init(train_x, train_y)

Initialize this model.

init(train_x: Array<Array<number>>, train_y: Array<(1 | -1)>)

Parameters

train_x (Array<Array<number>>) Training data

train_y (Array<(1 | -1)>) Target values

▸ fit()

Fit model.

fit()

▸ predict(data)

Returns predicted values.

predict(data: Array<Array<number>>): Array<number>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

SVR

Support vector regression

new SVR(kernel: any)

Parameters

kernel (any)

Instance Members

▸ init(x, y)

Initialize this model.

init(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ fit()

Fit model.

fit()

▸ predict(x)

lib/model/tietjen_moore.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

TheilSenRegression

lib/model/theil_sen.js

Theil-Sen regression

new TheilSenRegression()

Instance Members

▸ fit(x, y)

lib/model/theil_sen.js

Fit model.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(x)

lib/model/theil_sen.js

Returns predicted values.

predict(x: Array<number>): Array<number>

Parameters

x (Array<number>) Sample data

Returns

Array<number>: Predicted values

Thompson

lib/model/thompson.js

Thompson test

new Thompson(alpha: number)

Parameters

alpha (number) Significance level

Instance Members

▸ predict(data)

lib/model/thompson.js

Returns a list of the data predicted as outliers or not.

predict(data: Array<Array<number>>): Array<boolean>

Parameters

data (Array<Array<number>>) Training data

Returns

Array<boolean>: Predicted values

TietjenMoore

Tietjen-Moore Test

new TietjenMoore(k: number, threshold: number)

Parameters

k (number) Number of outliers

threshold (number) Threshold

Instance Members

▸ predict(data)

lib/model/tietjen_moore.js

Returns a list of the data predicted as outliers or not.

predict(data: Array<Array<number>>): Array<boolean>

Parameters

data (Array<Array<number>>) Training data

Returns

Array<boolean>: Predicted values

TighterPerceptron

lib/model/tighter_perceptron.js

Tighter Budget Perceptron

new TighterPerceptron(beta: number?, p: number?, update: ("perceptron" | "mira" | "nobias")?)

Parameters

beta (number? = 0) Margine

p (number? = 0) Cachs size

update (("perceptron" | "mira" | "nobias")? = 'perceptron') Update rule

Instance Members

▸ fit(x, y)

lib/model/tighter_perceptron.js

Fit model parameters.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

lib/model/tighter_perceptron.js

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

TightestPerceptron

lib/model/tightest_perceptron.js

Tightest Perceptron

new TightestPerceptron(b: number?, kernel: any, accuracyLoss: ("zero_one" | "hinge")?)

Parameters

b (number? = 10) Budget size

kernel (any = 'gaussian')

accuracyLoss (("zero_one" | "hinge")? = 'hinge') Accuracy loss type name

Instance Members

▸ fit(x, y)

lib/model/tightest_perceptron.js

Fit model parameters.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

lib/model/tightest_perceptron.js

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

TrigonometricInterpolation

lib/model/trigonometric_interpolation.js

Trigonometric interpolation

new TrigonometricInterpolation()

Instance Members

▸ fit(x, y)

lib/model/trigonometric_interpolation.js

Fit model.

fit(x: Array<number>, y: Array<number>)

Parameters

x (Array<number>) Training data

y (Array<number>) Target values

▸ predict(target)

lib/model/trigonometric_interpolation.js

Returns predicted interpolated values.

predict(target: Array<number>): Array<number>

Parameters

target (Array<number>) Sample data

Returns

Array<number>: Predicted values

SNE

Stochastic Neighbor Embedding

new SNE(datas: Array<Array<number>>, rd: number?, perplexity: number?)

Parameters

datas (Array<Array<number>>) Training data

rd (number? = 1) Reduced dimension

perplexity (number? = 30) Perplexity

Instance Members

▸ fit()

Fit model and returns reduced values.

fit(): Array<Array<number>>

Returns

Array<Array<number>>: Predicted values

▸ predict()

Returns reduced values.

predict(): Array<Array<number>>

Returns

Array<Array<number>>: Predicted values

tSNE

T-distributed Stochastic Neighbor Embedding

new tSNE(datas: Array<Array<number>>, rd: number?, perplexity: number?)

Parameters

datas (Array<Array<number>>) Training data

rd (number? = 1) Reduced dimension

perplexity (number? = 30) Perplexity

Instance Members

▸ fit()

Fit model and returns reduced values.

fit(): Array<Array<number>>

Returns

Array<Array<number>>: Predicted values

▸ predict()

lib/model/tukey_regression.js

Returns reduced values.

predict(): Array<Array<number>>

Returns

Array<Array<number>>: Predicted values

TukeyRegression

Tukey regression

new TukeyRegression(e: number)

Parameters

e (number) Error tolerance

Instance Members

▸ fit(x, y)

lib/model/tukey_regression.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ predict(x)

lib/model/tukey_regression.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

TukeysFences

lib/model/tukeys_fences.js

Tukey's fences

new TukeysFences(k: number)

Parameters

k (number) Tuning parameter

Instance Members

▸ predict(data)

lib/model/tukeys_fences.js

Returns a list of the data predicted as outliers or not.

predict(data: Array<Array<number>>): Array<boolean>

Parameters

data (Array<Array<number>>) Training data

Returns

Array<boolean>: Predicted values

RuLSIF

Relative unconstrained Least-Squares Importance Fitting

new RuLSIF(sigma: Array<number>, lambda: Array<number>, alpha: number, kernelNum: number)

Parameters

sigma (Array<number>) Sigmas of normal distribution

lambda (Array<number>) Regularization parameters

alpha (number) Relative parameter

kernelNum (number) Number of kernels

Instance Members

▸ fit(x1, x2)

Fit model.

fit(x1: Array<Array<number>>, x2: Array<Array<number>>)

Parameters

x1 (Array<Array<number>>) Numerator data

x2 (Array<Array<number>>) Denominator data

▸ predict(x)

Returns estimated values.

predict(x: Array<Array<number>>): Array<number>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

uLSIF

lib/model/universal_set_naive_bayes.js

unconstrained Least-Squares Importance Fitting

new uLSIF(sigma: Array<number>, lambda: Array<number>, kernelNum: number)

Extends RuLSIF

Parameters

sigma (Array<number>) Sigma of normal distribution

lambda (Array<number>) Regularization parameters

kernelNum (number) Number of kernels

UMAP

lib/model/umap.js

Uniform Manifold Approximation and Projection

new UMAP(datas: Array<Array<number>>, rd: number, n: number?, min_dist: number?)

Parameters

datas (Array<Array<number>>) Training data

rd (number) Reduced dimension

n (number? = 10) Number of neighborhoods

min_dist (number? = 0.1) Minimum distance

Instance Members

▸ fit()

lib/model/umap.js

Fit model and returns reduced values.

fit(): Array<Array<number>>

Returns

Array<Array<number>>: Predicted values

▸ predict()

lib/model/umap.js

Returns reduced values.

predict(): Array<Array<number>>

Returns

Array<Array<number>>: Predicted values

UniversalSetNaiveBayes

Universal-set Naive bayes

new UniversalSetNaiveBayes(distribution: "gaussian"?)

Parameters

distribution ("gaussian"? = 'gaussian') Distribution name

Instance Members

▸ fit(datas, labels)

lib/model/universal_set_naive_bayes.js

Fit model.

fit(datas: Array<Array<number>>, labels: Array<any>)

Parameters

datas (Array<Array<number>>) Training data

labels (Array<any>) Target values

▸ predict(data)

lib/model/universal_set_naive_bayes.js

Returns predicted categories.

predict(data: Array<Array<number>>): Array<any>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

VAE

Variational Autoencoder

new VAE(in_size: number, noise_dim: number, enc_layers: Array<LayerObject>, dec_layers: Array<LayerObject>, optimizer: string, class_size: (number | null), type: ("" | "conditional"))

Parameters

in_size (number) Input size

noise_dim (number) Number of noise dimension

enc_layers (Array<LayerObject>) Layers of encoder

dec_layers (Array<LayerObject>) Layers of decoder

optimizer (string) Optimizer of the network

class_size ((number | null)) Class size for conditional type

type (("" | "conditional")) Type name

Instance Members

▸ epoch

Epoch

epoch

Type: number

▸ fit(x, y, iteration, rate, batch)

Fit model.

fit(x: Array<Array<number>>, y: (Array<Array<number>> | null), iteration: number, rate: number, batch: number): number

Parameters

x (Array<Array<number>>) Training data

y ((Array<Array<number>> | null)) Conditional values

iteration (number) Iteration count

rate (number) Learning rate

batch (number) Batch size

Returns

number: Loss value

▸ predict(x, y)

Returns predicted values.

predict(x: Array<Array<number>>, y: (Array<Array<number>> | null)): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

y ((Array<Array<number>> | null)) Conditional values

Returns

Array<Array<number>>: Predicted values

▸ reduce(x, y)

Returns predicted values.

reduce(x: Array<Array<number>>, y: (Array<Array<number>> | null)): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

y ((Array<Array<number>> | null)) Conditional values

Returns

Array<Array<number>>: Predicted values

VAR

lib/model/var.js

Vector Autoregressive model

new VAR(p: number)

Parameters

p (number) Order

Instance Members

▸ fit(data)

lib/model/var.js

Fit model.

fit(data: Array<Array<number>>)

Parameters

data (Array<Array<number>>) Training data

▸ predict(data, k)

lib/model/var.js

Returns predicted future values.

predict(data: Array<Array<number>>, k: number): Array<Array<number>>

Parameters

data (Array<Array<number>>) Sample data

k (number) Prediction count

Returns

Array<Array<number>>: Predicted values

VBGMM

Variational Gaussian Mixture Model

new VBGMM(a: number, b: number, k: number)

Parameters

a (number) Tuning parameter

b (number) Tuning parameter

k (number) Initial number of clusters

Instance Members

▸ means

Means

means

Type: Matrix

▸ covs

Covariances

covs

Type: Array<Matrix>

▸ effectivity

Effectivity

effectivity

Type: Array<boolean>

▸ init(datas)

Initialize model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ fit()

Fit model.

fit()

▸ probability(data)

Returns probability of the datas.

probability(data: Array<Array<number>>): Matrix

Parameters

data (Array<Array<number>>) Sample data

Returns

Matrix: Predicted values

▸ predict(data)

lib/model/voted_perceptron.js

Returns predicted categories.

predict(data: Array<Array<number>>): Array<number>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

VotedPerceptron

Voted-perceptron

new VotedPerceptron(rate: number?)

Parameters

rate (number? = 1) Learning rate

Instance Members

▸ fit(x, y)

lib/model/voted_perceptron.js

Fit model.

fit(x: Array<Array<number>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<number>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

lib/model/voted_perceptron.js

Returns predicted values.

predict(data: Array<Array<number>>): Array<(1 | -1)>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

WeightedKMeans

Weighted k-means model

new WeightedKMeans(beta: number)

Parameters

beta (number) Tuning parameter

Instance Members

▸ centroids

Centroids

centroids

Type: Array<Array<number>>

▸ size

Number of clusters.

size

Type: number

▸ add(datas)

Add a new cluster.

add(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Training data

Returns

Array<number>: Added centroid

▸ clear()

Clear all clusters.

clear()

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

▸ fit(datas)

lib/model/weighted_knn.js

Fit model and returns total distance the centroid has moved.

fit(datas: Array<Array<number>>): number

Parameters

datas (Array<Array<number>>) Training data

Returns

number: Total distance the centroid has moved

WeightedKNN

Weighted K-Nearest Neighbor

Parameters

k (number) Number of neighbors

metric (("euclid" | "manhattan" | "chebyshev" | "minkowski" | function (Array<number>, Array<number>): number)? = 'euclid') Metric name

weight

(("gaussian" | "rectangular" | "triangular" | "epanechnikov" | "quartic" | "triweight" | "cosine" | "inversion")? = 'gaussian')

Weighting scheme name

Instance Members

▸ fit(x, y)

lib/model/weighted_knn.js

Fit model.

fit(x: Array<Array<number>>, y: Array<any>)

Parameters

x (Array<Array<number>>) Training data

y (Array<any>) Target values

▸ predict(data)

lib/model/weighted_knn.js

Returns predicted categories.

predict(data: Array<Array<number>>): Array<any>

Parameters

data (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

WeightedLeastSquares

lib/model/weighted_least_squares.js

Weighted least squares

new WeightedLeastSquares()

Instance Members

▸ fit(x, y, w)

lib/model/weighted_least_squares.js

Fit model.

fit(x: Array<Array<number>>, y: Array<Array<number>>, w: Array<number>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

w (Array<number>) Weight values

▸ predict(x)

lib/model/weighted_least_squares.js

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

Winnow

lib/model/winnow.js

Winnow

new Winnow(alpha: boolean?, threshold: number?, version: (1 | 2)?)

Parameters

alpha (boolean? = 2) Learning rate

threshold (number? = null) Threshold

version ((1 | 2)? = 1) Version of model

Instance Members

▸ fit(x, y)

lib/model/winnow.js

Fit model.

fit(x: Array<Array<(0 | 1)>>, y: Array<(1 | -1)>)

Parameters

x (Array<Array<(0 | 1)>>) Training data

y (Array<(1 | -1)>) Target values

▸ predict(data)

lib/model/winnow.js

Returns predicted values.

predict(data: Array<Array<(0 | 1)>>): Array<(1 | -1)>

Parameters

data (Array<Array<(0 | 1)>>) Sample data

Returns

Array<(1 | -1)>: Predicted values

Word2Vec

Word2Vec

new Word2Vec(method: ("CBOW" | "skip-gram"), n: number, wordsOrNumber: (number | Array<string>), reduce_size: number, optimizer: string)

Parameters

method (("CBOW" | "skip-gram")) Method name

n (number) Number of how many adjacent words to learn

wordsOrNumber ((number | Array<string>)) Initial words or number of words

reduce_size (number) Reduced dimension

optimizer (string) Optimizer of the network

Instance Members

▸ epoch

Epoch

epoch

Type: number

▸ fit(words, iteration, rate, batch)

Fit model.

fit(words: Array<string>, iteration: number, rate: number, batch: number): number

Parameters

words (Array<string>) Training data

iteration (number) Iteration count

rate (number) Learning rate

batch (number) Batch size

Returns

number: Loss value

▸ predict(x)

Returns predicted values.

predict(x: Array<string>): Array<Array<number>>

Parameters

x (Array<string>) Sample data

Returns

Array<Array<number>>: Predicted values

▸ reduce(x)

Returns reduced values.

reduce(x: Array<string>): Array<Array<number>>

Parameters

x (Array<string>) Sample data

Returns

Array<Array<number>>: Predicted values

XGBoost

eXtreme Gradient Boosting regression

new XGBoost(maxdepth: number?, srate: number?, lambda: number?, lr: number?)

Parameters

maxdepth (number? = 1) Maximum depth of tree

srate (number? = 1.0) Sampling rate

lambda (number? = 0.1) Regularization parameter

lr (number? = 0.5) Learning rate

Instance Members

▸ size

Number of trees

size

Type: number

▸ init(x, y)

Initialize model.

init(x: Array<Array<number>>, y: Array<Array<number>>)

Parameters

x (Array<Array<number>>) Training data

y (Array<Array<number>>) Target values

▸ fit()

Fit model.

fit()

▸ predict(x)

Returns predicted values.

predict(x: Array<Array<number>>): Array<Array<number>>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<Array<number>>: Predicted values

XGBoostClassifier

eXtreme Gradient Boosting classifier

new XGBoostClassifier(maxdepth: number?, srate: number?, lambda: number?, lr: number?)

Extends XGBoost

Parameters

maxdepth (number? = 1) Maximum depth of tree

srate (number? = 1.0) Sampling rate

lambda (number? = 0.1) Regularization parameter

lr (number? = 0) Learning rate

Instance Members

▸ init(x, y)

Initialize model.

init(x: Array<Array<number>>, y: Array<any>)

Parameters

x (Array<Array<number>>) Training data

y (Array<any>) Target values

▸ predict(x)

Returns predicted categories.

predict(x: Array<Array<number>>): Array<any>

Parameters

x (Array<Array<number>>) Sample data

Returns

Array<any>: Predicted values

XMeans

x-means

new XMeans()

Instance Members

▸ centroids

Centroids

centroids

Type: Array<Array<number>>

▸ size

Number of clusters.

size

Type: number

▸ clear()

Clear all clusters.

clear()

▸ fit(datas, iterations)

Fit model.

fit(datas: Array<Array<number>>, iterations: number)

Parameters

datas (Array<Array<number>>) Training data

iterations (number = -1) Iteration count

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

YeoJohnson

lib/model/yeo_johnson.js

Yeo-Johnson power transformation

new YeoJohnson(lambda: number?)

Parameters

lambda (number? = null) Lambda

Instance Members

▸ fit(x)

lib/model/yeo_johnson.js

Fit model.

fit(x: (Array<number> | Array<Array<number>>))

Parameters

x ((Array<number> | Array<Array<number>>)) Training data

▸ predict(x)

lib/model/yeo_johnson.js

Returns transformed values.

predict(x: (Array<number> | Array<Array<number>>)): (Array<number> | Array<Array<number>>)

Parameters

x ((Array<number> | Array<Array<number>>)) Sample data

Returns

(Array<number> | Array<Array<number>>): Predicted values

YinyangKMeans

Yinyang k-Means algorithm

new YinyangKMeans(k: number, t: number?)

Parameters

k (number) Number of clusters

t (number?) Number of groups

Instance Members

▸ centroids

Centroids

centroids

Type: Array<Array<number>>

▸ init(datas)

Initialize this model.

init(datas: Array<Array<number>>)

Parameters

datas (Array<Array<number>>) Training data

▸ fit()

Fit model.

fit()

▸ predict(datas)

Returns predicted categories.

predict(datas: Array<Array<number>>): Array<number>

Parameters

datas (Array<Array<number>>) Sample data

Returns

Array<number>: Predicted values

ZeroInflatedNegativeBinomial

lib/model/zinb.js

Zero-inflated negative binomial

new ZeroInflatedNegativeBinomial()

Instance Members

▸ fit(x)

lib/model/zinb.js

Fit model.

fit(x: Array<number>)

Parameters

x (Array<number>) Training data

▸ probability(x)

lib/model/zinb.js

Returns predicted probabilities.

probability(x: Array<number>): Array<number>

Parameters

x (Array<number>) Sample data

Returns

Array<number>: Predicted values

ZeroInflatedPoisson

lib/model/zip.js

Zero-inflated poisson

new ZeroInflatedPoisson(method: ("moments" | "maximum_likelihood")?)

Parameters

method (("moments" | "maximum_likelihood")? = 'maximum_likelihood') Method name

Instance Members

▸ fit(x)

lib/model/zip.js

Fit model.

fit(x: Array<number>)

Parameters

x (Array<number>) Training data

▸ probability(x)

lib/model/zip.js

Returns predicted probabilities.

probability(x: Array<number>): Array<number>

Parameters

x (Array<number>) Sample data

Returns

Array<number>: Predicted values

ZeroTruncatedPoisson

lib/model/ztp.js

Zero-truncated poisson

new ZeroTruncatedPoisson()

Instance Members

▸ fit(x)

lib/model/ztp.js

Fit model.

fit(x: Array<number>)

Parameters

x (Array<number>) Training data

▸ probability(x)

lib/model/ztp.js

Returns predicted probabilities.

probability(x: Array<number>): Array<number>

Parameters

x (Array<number>) Sample data

Returns

Array<number>: Predicted values

AcrobotRLEnvironment

lib/rl/acrobot.js

Acrobot environment

new AcrobotRLEnvironment()

Extends RLEnvironmentBase

RLRealRange

Real number range state/actioin

new RLRealRange(min: number, max: number)

Parameters

min (number) Minimum value

max (number) Maximum value

Instance Members

▸ toSpace(resolution)

Returns spaces.

toSpace(resolution: number): Array<number>

Parameters

resolution (number) Resolution value

Returns

Array<number>: Representative value

▸ toArray(resolution)

Returns array.

toArray(resolution: number): Array<number>

Parameters

resolution (number) Resolution value

Returns

Array<number>: Array of center values

▸ indexOf(value, resolution)

Returns index of the value.

indexOf(value: number, resolution: number): number

Parameters

value (number) Check value

resolution (number) Resolution value

Returns

number: Index of the value

RLIntRange

Integer number range state/actioin

new RLIntRange(min: number, max: number)

Parameters

min (number) Minimum value

max (number) Maximum value

Instance Members

▸ length

Length

length

Type: number

▸ toArray(resolution)

Returns array.

toArray(resolution: number): Array<number>

Parameters

resolution (number) Resolution value

Returns

Array<number>: Representative value

▸ indexOf(value, resolution)

Returns index of the value.

indexOf(value: number, resolution: number): number

Parameters

value (number) Check value

resolution (number) Resolution value

Returns

number: Index of the value

RLEnvironmentBase

Base class for reinforcement learning environment

new RLEnvironmentBase()

Properties

actions (Array<(Array<any> | RLRealRange | RLIntRange)>) : Action variables

states (Array<(Array<any> | RLRealRange | RLIntRange)>) : States variables

Instance Members

▸ epoch

Epoch

epoch

Type: number

▸ reward

Reward

reward

Parameters

value (object) Reward object

▸ clone()

Returns cloned environment.

clone(): RLEnvironmentBase

Returns

RLEnvironmentBase: Cloned environment

▸ close()

Close environment.

close()

▸ reset()

Reset environment.

reset()

▸ state(agent)

Returns current state.

state(agent: any): Array<any>

Parameters

agent (any) Agent

Returns

Array<any>: Current state

▸ setState(state, agent)

Set new state.

setState(state: Array<any>, agent: any)

Parameters

state (Array<any>) New state

agent (any) Agent

▸ step(action, agent)

Do action and returns new state.

step(action: Array<any>, agent: any)

Parameters

action (Array<any>) Actions to be performed by the agent

agent (any) Agent

▸ test(state, action, agent)

Do actioin without changing environment and returns new state.

test(state: Array<any>, action: Array<any>, agent: any)

Parameters

state (Array<any>) Environment state

action (Array<any>) Actions to be performed by the agent

agent (any) Agent

▸ sample_action(agent)

Sample an action.

sample_action(agent: any): Array<any>

Parameters

agent (any) Agent

Returns

Array<any>: Sampled action

EmptyRLEnvironment

Empty environment

new EmptyRLEnvironment()

Extends RLEnvironmentBase

BlackjackRLEnvironment

lib/rl/blackjack.js

Blackjack environment

new BlackjackRLEnvironment()

Extends RLEnvironmentBase

BreakerRLEnvironment

lib/rl/breaker.js

Breaker environment

new BreakerRLEnvironment()

Extends RLEnvironmentBase

CartPoleRLEnvironment

lib/rl/cartpole.js

Cartpole environment

new CartPoleRLEnvironment()

Extends RLEnvironmentBase

DraughtsRLEnvironment

lib/rl/draughts.js

Draughts environment

new DraughtsRLEnvironment()

Extends RLEnvironmentBase

GomokuRLEnvironment

lib/rl/gomoku.js

Gomoku environment

new GomokuRLEnvironment()

Extends RLEnvironmentBase

GridMazeRLEnvironment

lib/rl/grid.js

Grid world environment

new GridMazeRLEnvironment()

Extends RLEnvironmentBase

InHypercubeRLEnvironment

lib/rl/inhypercube.js

In-hypercube environment

new InHypercubeRLEnvironment(d: number?)

Extends RLEnvironmentBase

Parameters

d (number? = 2) Dimension of the environment

SmoothMazeRLEnvironment

lib/rl/maze.js

Smooth maze environment

new SmoothMazeRLEnvironment(width: number, height: number)

Extends RLEnvironmentBase

Parameters

width (number) Area width

height (number) Area height

MountainCarRLEnvironment

lib/rl/mountaincar.js

MountainCar environment

new MountainCarRLEnvironment()

Extends RLEnvironmentBase

PendulumRLEnvironment

lib/rl/pendulum.js

Pendulum environment

new PendulumRLEnvironment()

Extends RLEnvironmentBase

ReversiRLEnvironment

lib/rl/reversi.js

Reversi environment

new ReversiRLEnvironment()

Extends RLEnvironmentBase

WaterballRLEnvironment

lib/rl/waterball.js

Waterball environment

new WaterballRLEnvironment(width: number, height: number)

Extends RLEnvironmentBase

Parameters

width (number) Area width

height (number) Area height

accuracy

Returns accuracy.

accuracy(pred: Array<any>, t: Array<any>): number

Parameters

pred (Array<any>) Predicted classes

t (Array<any>) True classes

Returns

number: Accuracy

precision

Returns precision with macro average.

precision(pred: Array<any>, t: Array<any>): number

Parameters

pred (Array<any>) Predicted classes

t (Array<any>) True classes

Returns

number: Precision

recall

Returns recall with macro average.

recall(pred: Array<any>, t: Array<any>): number

Parameters

pred (Array<any>) Predicted classes

t (Array<any>) True classes

Returns

number: Recall

fScore

Returns F-score with macro average.

fScore(pred: Array<any>, t: Array<any>, beta: number?): number

Parameters

pred (Array<any>) Predicted classes

t (Array<any>) True classes

beta (number? = 1) Positive real factor. Recall is considered beta times as important as precision.

Returns

number: F-score

cohensKappa

Returns Cohen's kappa coefficient.

cohensKappa(pred: Array<any>, t: Array<any>): number

Parameters

pred (Array<any>) Predicted classes

t (Array<any>) True classes

Returns

number: Cohen's kappa coefficient

davisBouldinIndex

Returns Davies-Bouldin index.

davisBouldinIndex(data: Array<Array<number>>, pred: Array<any>, p: number, q: number): number

Parameters

data (Array<Array<number>>) Original data

pred (Array<any>) Predicted categories

p (number = 2) P

q (number = 1) Q

Returns

number: Davies-Bouldin index

silhouetteCoefficient

Returns Silhouette coefficient.

silhouetteCoefficient(data: Array<Array<number>>, pred: Array<any>): Array<number>

Parameters

data (Array<Array<number>>) Original data

pred (Array<any>) Predicted categories

Returns

Array<number>: Silhouette coefficient

dunnIndex

Returns Dunn index.

dunnIndex(data: Array<Array<number>>, pred: Array<any>, intra_d: ("max" | "mean" | "centroid"), inter_d: "centroid"): number

Parameters

data (Array<Array<number>>) Original data

pred (Array<any>) Predicted categories

intra_d (("max" | "mean" | "centroid") = 'max') Intra-cluster distance type

inter_d ("centroid" = 'centroid') Inter-cluster distance type

Returns

number: Dunn index

purity

Returns Purity.

purity(pred: Array<any>, t: Array<any>): number

Parameters

pred (Array<any>) Predicted categories

t (Array<any>) True categories

Returns

number: Purity

randIndex

Returns Rand index.

randIndex(pred: Array<any>, t: Array<any>): number

Parameters

pred (Array<any>) Predicted categories

t (Array<any>) True categories

Returns

number: Rank index

diceIndex

Returns Dice index.

diceIndex(pred: Array<any>, t: Array<any>, beta: number?): number

Parameters

pred (Array<any>) Predicted categories

t (Array<any>) True categories

beta (number? = 1) Positive real factor. Recall is considered beta times as important as precision.

Returns

number: Dice index

jaccardIndex

Returns Jaccard index.

jaccardIndex(pred: Array<any>, t: Array<any>): number

Parameters

pred (Array<any>) Predicted categories

t (Array<any>) True categories

Returns

number: Jaccard index

fowlkesMallowsIndex

lib/evaluate/dimensionality_reduction.js

Returns Fowlkes-Mallows index.

fowlkesMallowsIndex(pred: Array<any>, t: Array<any>): number

Parameters

pred (Array<any>) Predicted categories

t (Array<any>) True categories

Returns

number: Fowlkes-Mallows index

coRankingMatrix

Returns Co-Ranking Matrix.

coRankingMatrix(x: Array<Array<number>>, z: Array<Array<number>>, ks: number, kt: number): number

Parameters

x (Array<Array<number>>) Reduced values

z (Array<Array<number>>) Original values

ks (number) Rank significance

kt (number) Failure tolerance

Returns

number: Co-Ranking Matrix value

mse

Returns MSE (Mean Squared Error).

mse(pred: (Array<number> | Array<Array<number>>), t: (Array<number> | Array<Array<number>>)): (number | Array<number>)

Parameters

pred ((Array<number> | Array<Array<number>>)) Predicted values

t ((Array<number> | Array<Array<number>>)) True values

Returns

(number | Array<number>): Mean Squared Error

rmse

Returns RMSE (Root Mean Squared Error).

rmse(pred: (Array<number> | Array<Array<number>>), t: (Array<number> | Array<Array<number>>)): (number | Array<number>)

Parameters

pred ((Array<number> | Array<Array<number>>)) Predicted values

t ((Array<number> | Array<Array<number>>)) True values

Returns

(number | Array<number>): Root Mean Squared Error

mae

Returns MAE (Mean Absolute Error).

mae(pred: (Array<number> | Array<Array<number>>), t: (Array<number> | Array<Array<number>>)): (number | Array<number>)

Parameters

pred ((Array<number> | Array<Array<number>>)) Predicted values

t ((Array<number> | Array<Array<number>>)) True values

Returns

(number | Array<number>): Mean Absolute Error

mad

Returns MAD (Median Absolute Deviation).

mad(pred: (Array<number> | Array<Array<number>>), t: (Array<number> | Array<Array<number>>)): (number | Array<number>)

Parameters

pred ((Array<number> | Array<Array<number>>)) Predicted values

t ((Array<number> | Array<Array<number>>)) True values

Returns

(number | Array<number>): Median Absolute Deviation

rmspe

Returns RMSPE (Root Mean Squared Percentage Error).

rmspe(pred: (Array<number> | Array<Array<number>>), t: (Array<number> | Array<Array<number>>)): (number | Array<number>)

Parameters

pred ((Array<number> | Array<Array<number>>)) Predicted values

t ((Array<number> | Array<Array<number>>)) True values

Returns

(number | Array<number>): Root Mean Squared Percentage Error

mape

Returns MAPE (Mean Absolute Percentage Error).

mape(pred: (Array<number> | Array<Array<number>>), t: (Array<number> | Array<Array<number>>)): (number | Array<number>)

Parameters

pred ((Array<number> | Array<Array<number>>)) Predicted values

t ((Array<number> | Array<Array<number>>)) True values

Returns

(number | Array<number>): Mean Absolute Percentage Error

msle

Returns MSLE (Mean Squared Logarithmic Error).

msle(pred: (Array<number> | Array<Array<number>>), t: (Array<number> | Array<Array<number>>)): (number | Array<number>)

Parameters

pred ((Array<number> | Array<Array<number>>)) Predicted values

t ((Array<number> | Array<Array<number>>)) True values

Returns

(number | Array<number>): Mean Squared Logarithmic Error

rmsle

Returns RMSLE (Root Mean Squared Logarithmic Error).

rmsle(pred: (Array<number> | Array<Array<number>>), t: (Array<number> | Array<Array<number>>)): (number | Array<number>)

Parameters

pred ((Array<number> | Array<Array<number>>)) Predicted values

t ((Array<number> | Array<Array<number>>)) True values

Returns

(number | Array<number>): RootMean Squared Logarithmic Error

r2

Returns R2 (coefficient of determination).

r2(pred: (Array<number> | Array<Array<number>>), t: (Array<number> | Array<Array<number>>)): (number | Array<number>)

Parameters

pred ((Array<number> | Array<Array<number>>)) Predicted values

t ((Array<number> | Array<Array<number>>)) True values

Returns

(number | Array<number>): Coefficient of determination

correlation