weka.classifiers.trees.ht

Class NBNodeAdaptive

java.lang.Object

weka.classifiers.trees.ht.HNode

weka.classifiers.trees.ht.LeafNode

weka.classifiers.trees.ht.ActiveHNode

weka.classifiers.trees.ht.NBNode

weka.classifiers.trees.ht.NBNodeAdaptive

All Implemented Interfaces:

java.io.Serializable, LearningNode

public class NBNodeAdaptive
extends NBNode
implements LearningNode, java.io.Serializable

Implements a LearningNode that chooses between using majority class or naive Bayes for prediction

Version:

$Revision: 9705 $

Author:

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

See Also:

Serialized Form

Field Summary

Fields inherited from class weka.classifiers.trees.ht.ActiveHNode

m_weightSeenAtLastSplitEval

Fields inherited from class weka.classifiers.trees.ht.LeafNode

m_parentBranch, m_parentNode, m_theNode

Fields inherited from class weka.classifiers.trees.ht.HNode

m_classDistribution

Constructor Summary

Constructors

Constructor and Description
NBNodeAdaptive(Instances header, double nbWeightThreshold) Constructor

Method Summary

Modifier and Type	Method and Description
double[]	getDistribution(Instance inst, Attribute classAtt) Return a class probability distribution computed from the frequency counts at this node
void	updateNode(Instance inst) Update the node with the supplied instance

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

getPossibleSplits

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

classDistributionIsPure, graphTree, installNodeNums, isLeaf, leafForInstance, numEntriesInClassDistribution, toString, totalWeight, updateDistribution

Methods inherited from class java.lang.Object

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

Constructor Detail

NBNodeAdaptive

public NBNodeAdaptive(Instances header,
                      double nbWeightThreshold)
               throws java.lang.Exception

Constructor

Parameters:

header - the structure of the instances we're training from

nbWeightThreshold - the weight mass to see before allowing naive Bayes to predict

Throws:

java.lang.Exception - if a problem occurs

Method Detail

updateNode

public void updateNode(Instance inst)
                throws java.lang.Exception

Description copied from class: HNode

Update the node with the supplied instance

Overrides:

updateNode in class NBNode

Parameters:

inst - the instance to update with

Throws:

java.lang.Exception - if a problem occurs

getDistribution

public double[] getDistribution(Instance inst,
                                Attribute classAtt)
                         throws java.lang.Exception

Description copied from class: HNode

Return a class probability distribution computed from the frequency counts at this node

Overrides:

getDistribution in class NBNode

Parameters:

inst - the instance to get a prediction for

classAtt - the class attribute

Returns:

a class probability distribution

Throws:

java.lang.Exception - if a problem occurs