Class PLS

constructor

• new PLS(
      options: { latentVectors?: number; scale?: boolean; tolerance?: number },
      model: object,
  ): PLS

  Constructor for Partial Least Squares (PLS)

  Parameters

  • options: { latentVectors?: number; scale?: boolean; tolerance?: number }

    constructor options.

    • OptionallatentVectors?: number

      Number of latent vector to get (if the algorithm doesn't find a good model below the tolerance)

    • Optionalscale?: boolean

      rescale dataset using mean.

    • Optionaltolerance?: number

      tolerance used to stop the algorithm.

  • model: object

    for load purposes.

  Returns PLS

B

E

F

latentVectors

meanX

meanY

P

PBQ

Q

R2X

scale

scaleMethod

ssqYcal

stdDevX

stdDevY

T

tolerance

U

W

getExplainedVariance

• getExplainedVariance(): number

  Returns the explained variance on training of the PLS model

  Returns number

  the explained variance.

predict

• predict(dataset: any[] | Matrix): Matrix

  Predicts the behavior of the given dataset.

  Parameters

  • dataset: any[] | Matrix

    data to be predicted.

  Returns Matrix

  • predictions of each element of the dataset.

toJSON

• toJSON(): object

  Export the current model to JSON.

  Returns object

  • Current model.

train

• train(trainingSet: any[] | Matrix, trainingValues: any[] | Matrix): void

  Fits the model with the given data and predictions, in this function is calculated the following outputs:

  T - Score matrix of X P - Loading matrix of X U - Score matrix of Y Q - Loading matrix of Y B - Matrix of regression coefficient W - Weight matrix of X

  Parameters

  • trainingSet: any[] | Matrix

    matrix of features.

  • trainingValues: any[] | Matrix

    matrix of predictions.

  Returns void

Staticload

• load(model: object): PLS

  Load a PLS model from a JSON Object

  Parameters

  • model: object

    the serialized model to load.

  Returns PLS

  • PLS object from the given model