MOA 12.03
Real Time Analytics for Data Streams
moa.classifiers.core.splitcriteria.InfoGainSplitCriterion Class Reference

Class for computing splitting criteria using information gain with respect to distributions of class values. More...

Inheritance diagram for moa.classifiers.core.splitcriteria.InfoGainSplitCriterion:
Collaboration diagram for moa.classifiers.core.splitcriteria.InfoGainSplitCriterion:

List of all members.

Public Member Functions

double getMeritOfSplit (double[] preSplitDist, double[][] postSplitDists)
 Computes the merit of splitting for a given ditribution before the split and after it.
double getRangeOfMerit (double[] preSplitDist)
 Computes the range of splitting merit.
void getDescription (StringBuilder sb, int indent)
 Returns a string representation of this object.

Static Public Member Functions

static double computeEntropy (double[] dist)
static double computeEntropy (double[][] dists)
static int numSubsetsGreaterThanFrac (double[][] distributions, double minFrac)

Public Attributes

FloatOption minBranchFracOption

Protected Member Functions

void prepareForUseImpl (TaskMonitor monitor, ObjectRepository repository)
 This method describes the implementation of how to prepare this object for use.

Detailed Description

Class for computing splitting criteria using information gain with respect to distributions of class values.

The split criterion is used as a parameter on decision trees and decision stumps.

Author:
Richard Kirkby ([email protected])
Version:
Revision:
7

Definition at line 37 of file InfoGainSplitCriterion.java.


Member Function Documentation

static double moa.classifiers.core.splitcriteria.InfoGainSplitCriterion.computeEntropy ( double[]  dist) [static]
static double moa.classifiers.core.splitcriteria.InfoGainSplitCriterion.computeEntropy ( double  dists[][]) [static]

Definition at line 74 of file InfoGainSplitCriterion.java.

Here is the call graph for this function:

void moa.classifiers.core.splitcriteria.InfoGainSplitCriterion.getDescription ( StringBuilder  sb,
int  indent 
)

Returns a string representation of this object.

Used in AbstractMOAObject.toString to give a string representation of the object.

Parameters:
sbthe stringbuilder to add the description
indentthe number of characters to indent

Implements moa.MOAObject.

Definition at line 109 of file InfoGainSplitCriterion.java.

double moa.classifiers.core.splitcriteria.InfoGainSplitCriterion.getMeritOfSplit ( double[]  preSplitDist,
double  postSplitDists[][] 
)

Computes the merit of splitting for a given ditribution before the split and after it.

Parameters:
preSplitDistthe class distribution before the split
postSplitDistthe class distribution after the split
Returns:
value of the merit of splitting

Implements moa.classifiers.core.splitcriteria.SplitCriterion.

Definition at line 48 of file InfoGainSplitCriterion.java.

Here is the call graph for this function:

double moa.classifiers.core.splitcriteria.InfoGainSplitCriterion.getRangeOfMerit ( double[]  preSplitDist)

Computes the range of splitting merit.

Parameters:
preSplitDistthe class distribution before the split
Returns:
value of the range of splitting merit

Implements moa.classifiers.core.splitcriteria.SplitCriterion.

Definition at line 57 of file InfoGainSplitCriterion.java.

static int moa.classifiers.core.splitcriteria.InfoGainSplitCriterion.numSubsetsGreaterThanFrac ( double  distributions[][],
double  minFrac 
) [static]

Definition at line 88 of file InfoGainSplitCriterion.java.

Referenced by moa.classifiers.core.splitcriteria.InfoGainSplitCriterion.getMeritOfSplit().

Here is the caller graph for this function:

void moa.classifiers.core.splitcriteria.InfoGainSplitCriterion.prepareForUseImpl ( TaskMonitor  monitor,
ObjectRepository  repository 
) [protected, virtual]

This method describes the implementation of how to prepare this object for use.

All classes that extends this class have to implement prepareForUseImpl and not prepareForUse since prepareForUse calls prepareForUseImpl.

Parameters:
monitorthe TaskMonitor to use
repositorythe ObjectRepository to use

Implements moa.options.AbstractOptionHandler.

Definition at line 114 of file InfoGainSplitCriterion.java.


Member Data Documentation

Initial value:
 new FloatOption("minBranchFrac",
            'f',
            "Minimum fraction of weight required down at least two branches.",
            0.01, 0.0, 0.5)

Definition at line 42 of file InfoGainSplitCriterion.java.

Referenced by moa.classifiers.core.splitcriteria.InfoGainSplitCriterion.getMeritOfSplit().


The documentation for this class was generated from the following file:
 All Classes Namespaces Files Functions Variables Enumerations