ml-pls 4.3.2 | Documentation

ml-pls

4.3.2

OPLS ▸
- Static members
- .load
- Instance members
- #_getTrainTest
- #getLogs
- #predict
- #predictCategory
- #toJSON
PLS ▸
- Static members
- .load
- Instance members
- #getExplainedVariance
- #predict
- #toJSON
- #train
KOPLS ▸
- Static members
- .load
- Instance members
- #predict
- #toJSON
- #train
oplsNipals
tss

Need help reading this?

OPLS

src/OPLS.js

Creates new OPLS (orthogonal partial latent structures) from features and labels.

new OPLS(data: Array, labels: Array, options: Object)

Parameters

data (Array) matrix containing data (X).

labels (Array) 1D Array containing metadata (Y).

options

(Object
            = {})

Name	Description
options.center `boolean` (default `true`)	should the data be centered (subtract the mean).
options.scale `boolean` (default `true`)	should the data be scaled (divide by the standard deviation).
options.cvFolds `Array` (default `[]`)	Allows to provide folds as array of objects with the arrays trainIndex and testIndex as properties.
options.nbFolds `number` (default `7`)	Allows to generate the defined number of folds with the training and test set choosen randomly from the data set.

Static Members

▸ load(model)

src/OPLS.js

Load an OPLS model from JSON

load(model: Object): OPLS

Parameters

model (Object)

Returns

OPLS:

Instance Members

▸ _getTrainTest(X, group, index)

src/OPLS.js

_getTrainTest(X: any, group: any, index: any)

Parameters

X (any) dataset matrix object

group (any) labels matrix object

index (any) train and test index (output from getFold())

▸ getLogs()

src/OPLS.js

get access to all the computed elements Mainly for debug and testing

getLogs(): Object

Returns

Object: output object

▸ predict(features, options)

src/OPLS.js

Predict scores for new data

predict(features: Matrix, options: Object): Object

Parameters

features (Matrix) a matrix containing new data

options

(Object
            = {})

Name	Description
options.trueLabel `Array?`	an array with true values to compute confusion matrix
options.nc `Number?`	the number of components to be used

Returns

Object: predictions

▸ predictCategory(features, options)

src/OPLS.js

Predict scores for new data

predictCategory(features: Matrix, options: Object): Object

Parameters

features (Matrix) a matrix containing new data

options

(Object
            = {})

Name	Description
options.trueLabel `Array?`	an array with true values to compute confusion matrix
options.nc `Number?`	the number of components to be used

Returns

Object: predictions

▸ toJSON()

src/OPLS.js

Export the current model to a JSON object

toJSON(): Object

Returns

Object: model

PLS

src/PLS.js

new PLS()

Static Members

▸ load(model)

src/PLS.js

Load a PLS model from a JSON Object

load(model: object): PLS

Parameters

model (object)

Returns

PLS: PLS object from the given model

Instance Members

▸ getExplainedVariance()

src/PLS.js

Returns the explained variance on training of the PLS model

getExplainedVariance(): number

Returns

number:

▸ predict(dataset)

src/PLS.js

Predicts the behavior of the given dataset.

predict(dataset: (Matrix | Array)): Matrix

Parameters

dataset ((Matrix | Array)) data to be predicted.

Returns

Matrix: predictions of each element of the dataset.

▸ toJSON()

src/PLS.js

Export the current model to JSON.

toJSON(): object

Returns

object: Current model.

▸ train(trainingSet, trainingValues)

src/PLS.js

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

train(trainingSet: (Matrix | Array), trainingValues: (Matrix | Array))

Parameters

trainingSet ((Matrix | Array))

trainingValues ((Matrix | Array))

KOPLS

src/KOPLS.js

new KOPLS()

Static Members

▸ load(model, kernel)

src/KOPLS.js

Load a K-OPLS with the given model.

load(model: object, kernel: Kernel): KOPLS

Parameters

model (object)

kernel (Kernel) kernel used on the model, see ml-kernel .

Returns

KOPLS:

Instance Members

▸ predict(toPredict)

src/KOPLS.js

Predicts the output given the matrix to predict.

predict(toPredict: (Matrix | Array)): {y: Matrix, predScoreMat: Array<Matrix>, predYOrthVectors: Array<Matrix>}

Parameters

toPredict ((Matrix | Array))

Returns

{y: Matrix, predScoreMat: Array<Matrix>, predYOrthVectors: Array<Matrix>}: predictions

▸ toJSON()

src/KOPLS.js

Export the current model to JSON.

toJSON(): object

Returns

object: Current model.

▸ train(trainingSet, trainingValues)

src/KOPLS.js

Train the K-OPLS model with the given training set and labels.

train(trainingSet: (Matrix | Array), trainingValues: (Matrix | Array))

Parameters

trainingSet ((Matrix | Array))

trainingValues ((Matrix | Array))

oplsNipals

src/oplsNipals.js

OPLS loop

oplsNipals(data: (Array | Matrix), labels: (Array | Matrix), options: Object): Object

Parameters

data ((Array | Matrix)) matrix with features

labels ((Array | Matrix)) an array of labels (dependent variable)

options

(Object
            = {})

an object with options

Returns

Object: an object with model (filteredX: err, loadingsXOrtho: pOrtho, scoresXOrtho: tOrtho, weightsXOrtho: wOrtho, weightsPred: w, loadingsXpred: p, scoresXpred: t, loadingsY:)

tss

src/util/tss.js

Get total sum of square

tss(x: Array): Number

Parameters

x (Array) an array

Returns

Number: the sum of the squares