weka.classifiers.trees.ht

Class HNode

java.lang.Object

weka.classifiers.trees.ht.HNode

All Implemented Interfaces:

Serializable

Direct Known Subclasses:

LeafNode, SplitNode

public abstract class HNode
extends Object
implements Serializable

Abstract base class for nodes in a Hoeffding tree

Author:

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

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
Map<String,WeightMass>	m_classDistribution Class distribution at this node

Constructor Summary

Constructors

Constructor and Description
HNode() Construct a new HNode
HNode(Map<String,WeightMass> classDistrib) Construct a new HNode with the supplied class distribution

Method Summary

Methods

Modifier and Type	Method and Description
boolean	classDistributionIsPure() Returns true if the class distribution is pure
double[]	getDistribution(Instance inst, Attribute classAtt) Return a class probability distribution computed from the frequency counts at this node
void	graphTree(StringBuffer text)
int	installNodeNums(int nodeNum)
boolean	isLeaf() Returns true if this is a leaf
LeafNode	leafForInstance(Instance inst, SplitNode parent, String parentBranch) Return the leaf that the supplied instance ends up at
int	numEntriesInClassDistribution() The size of the class distribution
String	toString(boolean printLeaf) Print a textual description of the tree
double	totalWeight() Return the total weight of instances seen at this node
void	updateDistribution(Instance inst) Update the class frequency distribution with the supplied instance
abstract void	updateNode(Instance inst) Update the node with the supplied instance

Methods inherited from class java.lang.Object

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

Field Detail

m_classDistribution

public Map<String,WeightMass> m_classDistribution

Class distribution at this node

Constructor Detail

HNode

public HNode()

Construct a new HNode

HNode

public HNode(Map<String,WeightMass> classDistrib)

Construct a new HNode with the supplied class distribution

Parameters:

classDistrib -

Method Detail

isLeaf

public boolean isLeaf()

Returns true if this is a leaf

Returns:

numEntriesInClassDistribution

public int numEntriesInClassDistribution()

The size of the class distribution

Returns:

the number of entries in the class distribution

classDistributionIsPure

public boolean classDistributionIsPure()

Returns true if the class distribution is pure

Returns:

true if the class distribution is pure

updateDistribution

public void updateDistribution(Instance inst)

Update the class frequency distribution with the supplied instance

Parameters:

inst - the instance to update with

getDistribution

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

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

Parameters:

inst - the instance to get a prediction for

classAtt - the class attribute

Returns:

a class probability distribution

Throws:

Exception - if a problem occurs

installNodeNums

public int installNodeNums(int nodeNum)

graphTree

public void graphTree(StringBuffer text)

toString

public String toString(boolean printLeaf)

Print a textual description of the tree

Parameters:

printLeaf - true if leaf models (NB, NB adaptive) should be output

Returns:

a textual description of the tree

totalWeight

public double totalWeight()

Return the total weight of instances seen at this node

Returns:

the total weight of instances seen at this node

leafForInstance

public LeafNode leafForInstance(Instance inst,
                       SplitNode parent,
                       String parentBranch)

Return the leaf that the supplied instance ends up at

Parameters:

inst - the instance to find the leaf for

parent - the parent node

parentBranch - the parent branch

Returns:

the leaf that the supplied instance ends up at

updateNode

public abstract void updateNode(Instance inst)
                         throws Exception

Update the node with the supplied instance

Parameters:

inst - the instance to update with

Throws:

Exception - if a problem occurs