com.rapidminer.operator.learner.tree
Class InfoGainCriterion

java.lang.Object
  com.rapidminer.operator.learner.tree.AbstractCriterion
      com.rapidminer.operator.learner.tree.InfoGainCriterion

All Implemented Interfaces:

Criterion, MinimalGainHandler

Direct Known Subclasses:

GainRatioCriterion

public class InfoGainCriterion
extends AbstractCriterion
implements MinimalGainHandler

This criterion implements the well known information gain in order to calculate the benefit of a split. The information gain is defined as the change in entropy from a prior state to a state that takes some information as given by the entropy.

Author:

Sebastian Land, Ingo Mierswa

Field Summary

Fields inherited from class com.rapidminer.operator.learner.tree.AbstractCriterion

labelAttribute, leftLabelWeights, leftWeight, rightLabelWeights, rightWeight, totalLabelWeights, totalWeight, weightAttribute

Constructor Summary

InfoGainCriterion()

InfoGainCriterion(double minimalGain)

Method Summary

protected double	getBenefit(double[][] weightCounts)
protected double	getEntropy(double[] labelWeights, double totalWeight)
double	getIncrementalBenefit()
double	getNominalBenefit(ExampleSet exampleSet, Attribute attribute)
double	getNumericalBenefit(ExampleSet exampleSet, Attribute attribute, double splitValue)
void	setMinimalGain(double minimalGain)
boolean	supportsIncrementalCalculation()

Methods inherited from class com.rapidminer.operator.learner.tree.AbstractCriterion

startIncrementalCalculation, swapExample

Methods inherited from class java.lang.Object

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

Constructor Detail

InfoGainCriterion

public InfoGainCriterion()

InfoGainCriterion

public InfoGainCriterion(double minimalGain)

Method Detail

setMinimalGain

public void setMinimalGain(double minimalGain)

Specified by:

setMinimalGain in interface MinimalGainHandler

getNominalBenefit

public double getNominalBenefit(ExampleSet exampleSet,
                                Attribute attribute)

Specified by:

getNominalBenefit in interface Criterion

getNumericalBenefit

public double getNumericalBenefit(ExampleSet exampleSet,
                                  Attribute attribute,
                                  double splitValue)

Specified by:

getNumericalBenefit in interface Criterion

getBenefit

protected double getBenefit(double[][] weightCounts)

getEntropy

protected double getEntropy(double[] labelWeights,
                            double totalWeight)

supportsIncrementalCalculation

public boolean supportsIncrementalCalculation()

Specified by:

supportsIncrementalCalculation in interface Criterion

Overrides:

supportsIncrementalCalculation in class AbstractCriterion

getIncrementalBenefit

public double getIncrementalBenefit()

Specified by:

getIncrementalBenefit in interface Criterion

Overrides:

getIncrementalBenefit in class AbstractCriterion