Class uLSIF

unconstrained Least-Squares Importance Fitting

Hierarchy (View Summary)

RuLSIF
- uLSIF

Index

Constructors

Properties

_alpha _centers _kernelNum _kw _lambda _lambda_cand _sigma _sigma_cand

Methods

_kernel_gaussian fit predict

Constructors

constructor

new uLSIF(sigma: number[], lambda: number[], kernelNum: number): uLSIF
Parameters
- sigma: number[]
  Sigma of normal distribution
- lambda: number[]
  Regularization parameters
- kernelNum: number
  Number of kernels
Returns uLSIF
Overrides RuLSIF.constructor
- Defined in lib/model/ulsif.d.ts:43

Properties

_alpha

_alpha: number

_centers

_centers: any

_kernelNum

_kernelNum: number

_kw

_kw: Matrix<number>

_lambda

_lambda: number

_lambda_cand

_lambda_cand: number[]

_sigma

_sigma: number

_sigma_cand

_sigma_cand: number[]

Methods

_kernel_gaussian

_kernel_gaussian(x: any, c: any, s: any): Matrix<number[]>
Parameters
- x: any
- c: any
- s: any
Returns Matrix<number[]>
Inherited from RuLSIF._kernel_gaussian
- Defined in lib/model/ulsif.d.ts:16

fit

fit(x1: number[][], x2: number[][]): void
Fit model.
Parameters
- x1: number[][]
  Numerator data
- x2: number[][]
  Denominator data
Returns void
Inherited from RuLSIF.fit
- Defined in lib/model/ulsif.d.ts:22

predict

predict(x: number[][]): number[]
Returns estimated values.
Parameters
- x: number[][]
  Sample data
Returns number[]
Predicted values
Inherited from RuLSIF.predict
- Defined in lib/model/ulsif.d.ts:32