Class TSNE

A class for performing dimensionality reduction on a matrix of data.

Hierarchy

Constructors

constructor

Properties

_data _gains _initialized _iter _params _projection _randomizer _result _yStep

Accessors

dimensionality metric parameters seed

Methods

checkInit generator init next transform

Constructors

constructor

  • t-SNE algorithm for dimensionality reduction.

    Parameters

    • data: Matrix | number[][]

      A 2D array or matrix whose dimensionality is to be reduced.

    • Optional params: Partial<TSNEParams>

      Optional parameters for the algorithm.

    Returns TSNE

Properties

_data

_data: Matrix

Protected _gains

_gains: Matrix

Protected _initialized

_initialized: boolean = false

Protected _iter

_iter: number = 0

Protected _params

_params: TSNEParams & DimRedParams

_projection

_projection: Matrix

Protected _randomizer

_randomizer: Randomizer

_result

_result: Matrix

Protected _yStep

_yStep: Matrix

Accessors

dimensionality

  • get dimensionality(): number

  • Gets the dimensionality of the data after dimensionality reduction.

    Returns number

metric

  • get metric(): string

  • Gets the metric used for calculating distances between data points.

    Returns string

parameters

  • get parameters(): Omit<P & DimRedParams, "dimensionality" | "metric" | "seed">

  • Gets the parameters used for the algorithm.

    Returns Omit<P & DimRedParams, "dimensionality" | "metric" | "seed">

seed

  • get seed(): number

  • Gets the seed used for generating random numbers.

    Returns number

Methods

Protected checkInit

generator

  • A generator function that yields the projection of the data after each iteration.

    Parameters

    • iterations: number = 500

      The number of iterations to perform. Default to 500.

    Returns Generator<Matrix, Matrix, unknown>

    The projection of the data after dimensionality reduction.

    Yields

    The projection of the data after each iteration.

init

  • Initializes the t-SNE algorithm by computing pairwise distances between data points and computing probabilities for each pair of points. It also initializes the step and gains matrices.

    Returns TSNE

    The t-SNE instance.

Protected next

transform

  • Transforms the data by performing dimensionality reduction on it.

    Parameters

    • iterations: number = 500

      The number of iterations to perform. Default to 500.

    Returns Matrix

    The projection of the data after dimensionality reduction.

On This Page