org.apache.ignite.ml.trainers

Class TrainerTransformers

java.lang.Object

org.apache.ignite.ml.trainers.TrainerTransformers

public class TrainerTransformers
extends Object

Class containing various trainer transformers.

Constructor Summary

Constructors

Constructor and Description
TrainerTransformers()

Method Summary

Modifier and Type	Method and Description
static int[]	getMapping(int featuresVectorSize, int maximumFeaturesCntPerMdl, long seed) Get mapping R^featuresVectorSize -> R^maximumFeaturesCntPerMdl.
static <L> BaggedTrainer<L>	makeBagged(DatasetTrainer<? extends IgniteModel,L> trainer, int ensembleSize, double subsampleRatio, PredictionsAggregator aggregator) Add bagging logic to a given trainer.
static <M extends IgniteModel<Vector,Double>,L> BaggedTrainer<L>	makeBagged(DatasetTrainer<M,L> trainer, int ensembleSize, double subsampleRatio, int featureVectorSize, int featuresSubspaceDim, PredictionsAggregator aggregator) Add bagging logic to a given trainer.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

TrainerTransformers

public TrainerTransformers()

Method Detail

makeBagged

public static <L> BaggedTrainer<L> makeBagged(DatasetTrainer<? extends IgniteModel,L> trainer,
                                              int ensembleSize,
                                              double subsampleRatio,
                                              PredictionsAggregator aggregator)

Add bagging logic to a given trainer. No features bootstrapping is done.

Type Parameters:

L - Type of labels.

Parameters:

ensembleSize - Size of ensemble.

subsampleRatio - Subsample ratio to whole dataset.

aggregator - Aggregator.

Returns:

Bagged trainer.

makeBagged

public static <M extends IgniteModel<Vector,Double>,L> BaggedTrainer<L> makeBagged(DatasetTrainer<M,L> trainer,
                                                                                   int ensembleSize,
                                                                                   double subsampleRatio,
                                                                                   int featureVectorSize,
                                                                                   int featuresSubspaceDim,
                                                                                   PredictionsAggregator aggregator)

Add bagging logic to a given trainer.

Type Parameters:

M - Type of one model in ensemble.

L - Type of labels.

Parameters:

ensembleSize - Size of ensemble.

subsampleRatio - Subsample ratio to whole dataset.

aggregator - Aggregator.

featureVectorSize - Feature vector dimensionality.

featuresSubspaceDim - Feature subspace dimensionality.

Returns:

Bagged trainer.

getMapping

public static int[] getMapping(int featuresVectorSize,
                               int maximumFeaturesCntPerMdl,
                               long seed)

Get mapping R^featuresVectorSize -> R^maximumFeaturesCntPerMdl.

Parameters:

featuresVectorSize - Features vector size (Dimension of initial space).

maximumFeaturesCntPerMdl - Dimension of target space.

seed - Seed.

Returns:

Mapping R^featuresVectorSize -> R^maximumFeaturesCntPerMdl.