Package weka.classifiers.trees.ht

Class ConditionalSufficientStats

java.lang.Object

weka.classifiers.trees.ht.ConditionalSufficientStats

All Implemented Interfaces:

Serializable

Direct Known Subclasses:

GaussianConditionalSufficientStats, NominalConditionalSufficientStats

public abstract class ConditionalSufficientStats
extends Object
implements Serializable

Records sufficient stats for an attribute

Version:

$Revision: 9705 $

Author:

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

See Also:

• Serialized Form

Constructor Summary

Constructors

Constructor	Description
ConditionalSufficientStats()

Method Summary

Modifier and Type	Method	Description
abstract SplitCandidate	bestSplit(SplitMetric splitMetric, Map<String,WeightMass> preSplitDist, String attName)	Return the best split
abstract double	probabilityOfAttValConditionedOnClass(double attVal, String classVal)	Return the probability of an attribute value conditioned on a class value
abstract void	update(double attVal, String classVal, double weight)	Update this stat with the supplied attribute value and class value

Methods inherited from class java.lang.Object

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

Constructor Details

ConditionalSufficientStats

public ConditionalSufficientStats()

Method Details

update

public abstract void update(double attVal,
 String classVal,
 double weight)

Update this stat with the supplied attribute value and class value

Parameters:

attVal - the value of the attribute

classVal - the class value

weight - the weight of this observation

probabilityOfAttValConditionedOnClass

public abstract double probabilityOfAttValConditionedOnClass(double attVal,
 String classVal)

Return the probability of an attribute value conditioned on a class value

Parameters:

attVal - the attribute value to compute the conditional probability for

classVal - the class value

Returns:

the probability

bestSplit

public abstract SplitCandidate bestSplit(SplitMetric splitMetric,
 Map<String,WeightMass> preSplitDist,
 String attName)

Return the best split

Parameters:

splitMetric - the split metric to use

preSplitDist - the distribution of class values prior to splitting

attName - the name of the attribute being considered for splitting

Returns:

the best split for the attribute