org.apache.ignite.ml.composition.boosting

Class GDBRegressionTrainer

java.lang.Object

org.apache.ignite.ml.trainers.DatasetTrainer<ModelsComposition,Double>

org.apache.ignite.ml.composition.boosting.GDBTrainer

org.apache.ignite.ml.composition.boosting.GDBRegressionTrainer

Direct Known Subclasses:

GDBRegressionOnTreesTrainer

public abstract class GDBRegressionTrainer
extends GDBTrainer

Trainer for regressor using Gradient Boosting. This algorithm uses gradient of Mean squared error loss metric [MSE] in each step of learning.

Nested Class Summary

Nested classes/interfaces inherited from class org.apache.ignite.ml.composition.boosting.GDBTrainer

GDBTrainer.GDBModel

Nested classes/interfaces inherited from class org.apache.ignite.ml.trainers.DatasetTrainer

DatasetTrainer.EmptyDatasetException

Field Summary

Fields inherited from class org.apache.ignite.ml.composition.boosting.GDBTrainer

checkConvergenceStgyFactory, loss

Fields inherited from class org.apache.ignite.ml.trainers.DatasetTrainer

environment

Constructor Summary

Constructors

Constructor and Description
GDBRegressionTrainer(double gradStepSize, Integer cntOfIterations) Constructs instance of GDBRegressionTrainer.

Method Summary

Modifier and Type	Method and Description
protected double	externalLabelToInternal(double x) Maps external representation of label to internal.
protected double	internalLabelToExternal(double x) Maps internal representation of label to external.
protected <V,K> boolean	learnLabels(DatasetBuilder<K,V> builder, IgniteBiFunction<K,V,Vector> featureExtractor, IgniteBiFunction<K,V,Double> lExtractor) Defines unique labels in dataset if need (useful in case of classification).

Methods inherited from class org.apache.ignite.ml.composition.boosting.GDBTrainer

buildBaseModelTrainer, checkState, computeInitialValue, fit, getLearningStrategy, updateModel, withCheckConvergenceStgyFactory

Methods inherited from class org.apache.ignite.ml.trainers.DatasetTrainer

fit, fit, fit, fit, getLastTrainedModelOrThrowEmptyDatasetException, setEnvironment, update, update, update, update, update

Methods inherited from class java.lang.Object

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

Constructor Detail

GDBRegressionTrainer

public GDBRegressionTrainer(double gradStepSize,
                            Integer cntOfIterations)

Constructs instance of GDBRegressionTrainer.

Parameters:

gradStepSize - Grad step size.

cntOfIterations - Count of learning iterations.

Method Detail

learnLabels

protected <V,K> boolean learnLabels(DatasetBuilder<K,V> builder,
                                    IgniteBiFunction<K,V,Vector> featureExtractor,
                                    IgniteBiFunction<K,V,Double> lExtractor)

Defines unique labels in dataset if need (useful in case of classification).

Specified by:

learnLabels in class GDBTrainer

Parameters:

builder - Dataset builder.

featureExtractor - Feature extractor.

lExtractor - Labels extractor.

Returns:

true if labels learning was successful.

externalLabelToInternal

protected double externalLabelToInternal(double x)

Maps external representation of label to internal.

Specified by:

externalLabelToInternal in class GDBTrainer

Parameters:

x - Label value.

internalLabelToExternal

protected double internalLabelToExternal(double x)

Maps internal representation of label to external.

Specified by:

internalLabelToExternal in class GDBTrainer

Parameters:

x - Label value.