Package weka.classifiers.trees.ht

Class GiniSplitMetric

java.lang.Object

weka.classifiers.trees.ht.SplitMetric

weka.classifiers.trees.ht.GiniSplitMetric

All Implemented Interfaces:

Serializable

public class GiniSplitMetric
extends SplitMetric
implements Serializable

Implements the gini splitting criterion

Version:

$Revision$

Author:

Richard Kirkby (rkirkby@cs.waikato.ac.nz), Mark Hall (mhall{[at]}pentaho{[dot]}com)

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor	Description
GiniSplitMetric()

Method Summary

Modifier and Type	Method	Description
double	evaluateSplit(Map<String,WeightMass> preDist, List<Map<String,WeightMass>> postDist)	Evaluate the merit of a split
double	getMetricRange(Map<String,WeightMass> preDist)	Get the range of the splitting metric
static double	gini(Map<String,WeightMass> dist)	Return the gini metric computed from the supplied distribution

Methods inherited from class weka.classifiers.trees.ht.SplitMetric

sum

Methods inherited from class java.lang.Object

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

Constructor Detail

GiniSplitMetric

public GiniSplitMetric()

Method Detail

evaluateSplit

public double evaluateSplit(Map<String,WeightMass> preDist,
                            List<Map<String,WeightMass>> postDist)

Description copied from class: SplitMetric

Evaluate the merit of a split

Specified by:

evaluateSplit in class SplitMetric

Parameters:

preDist - the class distribution before the split

postDist - the class distributions after the split

Returns:

the merit of the split

gini

public static double gini(Map<String,WeightMass> dist)

Return the gini metric computed from the supplied distribution

Parameters:

dist - dist the distribution to compute the gini metric from

Returns:

getMetricRange

public double getMetricRange(Map<String,WeightMass> preDist)

Description copied from class: SplitMetric

Get the range of the splitting metric

Specified by:

getMetricRange in class SplitMetric

Parameters:

preDist - the pre-split class distribution

Returns:

the range of the splitting metric