weka.classifiers.trees.j48
Class InfoGainSplitCrit

java.lang.Object
  weka.classifiers.trees.j48.SplitCriterion
      weka.classifiers.trees.j48.EntropyBasedSplitCrit
          weka.classifiers.trees.j48.InfoGainSplitCrit

All Implemented Interfaces:

java.io.Serializable

public final class InfoGainSplitCrit

extends EntropyBasedSplitCrit

Class for computing the information gain for a given distribution.

Version:

$Revision: 1.6 $

Author:

Eibe Frank (eibe@cs.waikato.ac.nz)

See Also:

Serialized Form

Field Summary

Fields inherited from class weka.classifiers.trees.j48.EntropyBasedSplitCrit

log2

Constructor Summary

InfoGainSplitCrit()

Method Summary

double	splitCritValue(Distribution bags) This method is a straightforward implementation of the information gain criterion for the given distribution.
double	splitCritValue(Distribution bags, double totalNoInst) This method computes the information gain in the same way C4.5 does.
double	splitCritValue(Distribution bags, double totalNoInst, double oldEnt) This method computes the information gain in the same way C4.5 does.

Methods inherited from class weka.classifiers.trees.j48.EntropyBasedSplitCrit

logFunc, newEnt, oldEnt, splitEnt

Methods inherited from class weka.classifiers.trees.j48.SplitCriterion

splitCritValue, splitCritValue, splitCritValue

Methods inherited from class java.lang.Object

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

Constructor Detail

InfoGainSplitCrit

public InfoGainSplitCrit()

Method Detail

splitCritValue

public final double splitCritValue(Distribution bags)

This method is a straightforward implementation of the information gain criterion for the given distribution.

Overrides:

splitCritValue in class SplitCriterion

Returns:

value of splitting criterion. 0 by default

splitCritValue

public final double splitCritValue(Distribution bags,
                                   double totalNoInst)

This method computes the information gain in the same way C4.5 does.

Parameters:

totalNoInst - weight of ALL instances (including the ones with missing values).

splitCritValue

public final double splitCritValue(Distribution bags,
                                   double totalNoInst,
                                   double oldEnt)

This method computes the information gain in the same way C4.5 does.

Parameters:

totalNoInst - weight of ALL instances

oldEnt - entropy with respect to "no-split"-model.