svm.SVR

Linear Support Vector Regression. Similar to SVR with parameter kernel=’linear’, but implemented in terms of liblinear rather than libsvm, so it has more flexibility in the choice of penalties and loss functions and should scale better to large numbers of samples.

This class supports both dense and sparse input.

Constructors

Methods

Constructors


constructor

new SVR(options: `object`)

Defined in svm/classes.ts:254

Parameters:

ParamTypeDefaultDescription
options.cacheSizenumberCache size in MB
options.coef0numbercoef0 parameter for Polynomial and Sigmoid kernels
options.costnumberCost parameter, for C SVC, Epsilon SVR and NU SVR
options.degreenumberDegree of polynomial, test for polynomial kernel
options.epsilonnumberFor epsilon SVR
options.gammaGamma parameter of the RBF, Polynomial and Sigmoid kernels. Default value is 1/numfeatures
options.kernelKernelType of Kernel
options.nunumberFor NU SVC and NU SVR
options.probabilityEstimatesbooleanweather to train SVC/SVR model for probability estimates,
options.quietbooleanPrint info during training if false (aka verbose)
options.shrinkingbooleanUse shrinking euristics (faster)
options.tolerancenumberTolerance
options.weightSet weight for each possible class

Returns: SVR

Methods


λ fit

Fit the model according to the given training data.

Defined in svm/classes.ts:101

Parameters:

ParamTypeDefaultDescription
Xnumber[][]
ynumber[]

Returns:

🤘 Promise<void>

λ fromJSON

Restores the model from a JSON checkpoint

Defined in svm/classes.ts:158

Parameters:

ParamTypeDefaultDescription
options.optionsanynull
options.svmanynull
options.typeanynull

Returns:

void

λ predict

Predict using the linear model

Defined in svm/classes.ts:125

Parameters:

ParamTypeDefaultDescription
Xnumber[][]

Returns:

number[]

λ predictOne

Predict the label of one sample.

Defined in svm/classes.ts:135

Parameters:

ParamTypeDefaultDescription
Xnumber[]

Returns:

number[]

λ toJSON

Saves the current SVM as a JSON object

Defined in svm/classes.ts:144

Returns:

ParamTypeDescription
svmanyundefined