Class HMM

Hidden Markov model

Hierarchy

  • HMMBase
    • HMM

Constructors

constructor

Properties

_a _b _n _p _type _x_cand

Methods

_backward _bt _forward _gamma _update _xi bestPath fit probability

Constructors

constructor

  • new HMM(n: number): HMM

  • Parameters

    • n: number

      Number of states

    Returns HMM

Properties

_a

_a: Matrix

_b

_b: any

_n

_n: number

_p

_p: Matrix<T>

_type

_type: string

_x_cand

_x_cand: any[]

Methods

_backward

  • _backward(x: any, c?: any, prob?: boolean): Matrix[]

  • Parameters

    • x: any
    • Optionalc: any
    • Optionalprob: boolean

    Returns Matrix[]

_bt

  • _bt(x: any, t: any): Matrix<T>

  • Parameters

    • x: any
    • t: any

    Returns Matrix<T>

_forward

  • _forward(x: any, scaled?: boolean): (Matrix<T> | Matrix<T>[])[]

  • Parameters

    • x: any
    • Optionalscaled: boolean

    Returns (Matrix<T> | Matrix<T>[])[]

_gamma

  • _gamma(alpha: any, beta: any): Matrix[]

  • Parameters

    • alpha: any
    • beta: any

    Returns Matrix[]

_update

  • _update(gamma: any, xi: any): void

  • Parameters

    • gamma: any
    • xi: any

    Returns void

_xi

  • _xi(x: any, alpha: any, beta: any, c: any): any[][]

  • Parameters

    • x: any
    • alpha: any
    • beta: any
    • c: any

    Returns any[][]

bestPath

  • bestPath(data: any[][]): number[][]

  • Returns best path of the datas.

    Parameters

    • data: any[][]

      Sample data

    Returns number[][]

    Predicted path

fit

  • fit(datas: any[][], scaled?: boolean): void

  • Fit model.

    Parameters

    • datas: any[][]

      Training data

    • Optionalscaled: boolean

      Do scaled calculation or not

    Returns void

probability

  • probability(datas: any[][]): number[]

  • Returns probability of the datas.

    Parameters

    • datas: any[][]

      Sample data

    Returns number[]

    Predicted values

Settings

Member Visibility

On This Page

Constructors
Properties
Methods