Package weka.classifiers.trees

Class RandomRegressionForest.Node

java.lang.Object

weka.classifiers.trees.RandomRegressionForest.Node

All Implemented Interfaces:

Serializable

Enclosing class:

RandomRegressionForest

public class RandomRegressionForest.Node
extends Object
implements Serializable

TODO: description of class

Version:

$Revision$

Author:

Bernhard Pfahringer ([email protected])

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field	Description
protected RandomRegressionForest.Node	m_Less
protected LinearRegressionJ	m_LinearReg
protected RandomRegressionForest.Node	m_More
protected int	m_SplitIndex
protected double	m_SplitValue

Constructor Summary

Constructors

Constructor	Description
Node(weka.core.Instances data, Random r, int min)	the constructor

Method Summary

Modifier and Type	Method	Description
double	classifyInstance(weka.core.Instance instance)	classifies the given instance
void	findRandomSplit(weka.core.Instances data, Random r, int min)	determines a random split for the data, tries 10 pairs.
void	prefix(int indent, StringBuffer sb)	generates the tree structure prefix
boolean	subsetSizesOK(weka.core.Instances data, int min)	tests whether the leaf threshold is OK
void	toString(int indent, StringBuffer sb, List<String> models)	Generates a string representation of the node.
void	turnIntoLeaf(weka.core.Instances data)	turns the node into a leaf

Methods inherited from class java.lang.Object

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

Field Detail

m_SplitValue

protected double m_SplitValue

m_SplitIndex

protected int m_SplitIndex

m_LinearReg

protected LinearRegressionJ m_LinearReg

m_Less

protected RandomRegressionForest.Node m_Less

m_More

protected RandomRegressionForest.Node m_More

Constructor Detail

Node

public Node(weka.core.Instances data,
            Random r,
            int min)
     throws Exception

the constructor

Parameters:

data - the data to use

r - the random number generator to use

min - the leaf threshold

Throws:

Exception - if something goes wrong

Method Detail

turnIntoLeaf

public void turnIntoLeaf(weka.core.Instances data)
                  throws Exception

turns the node into a leaf

Parameters:

data - the data to use for linear regression

Throws:

Exception - if training of LinearRegression fails

classifyInstance

public double classifyInstance(weka.core.Instance instance)
                        throws Exception

classifies the given instance

Parameters:

instance - the instance to classify

Returns:

the regression value

Throws:

Exception - if the classification fails

findRandomSplit

public void findRandomSplit(weka.core.Instances data,
                            Random r,
                            int min)

determines a random split for the data, tries 10 pairs.

Parameters:

data - the data to use

r - the random number generator for

min - the leaf threshold

See Also:

m_SplitIndex, m_SplitValue

subsetSizesOK

public boolean subsetSizesOK(weka.core.Instances data,
                             int min)

tests whether the leaf threshold is OK

Parameters:

data - the data to use

min - the leaf threshold

Returns:

true if the size is OK

prefix

public void prefix(int indent,
                   StringBuffer sb)

generates the tree structure prefix

Parameters:

indent - the depth

sb - the StringBuffer to add the prefix to

toString

public void toString(int indent,
                     StringBuffer sb,
                     List<String> models)

Generates a string representation of the node.

Parameters:

indent - the depth

sb - the StringBuffer to add the output to

models - the list to add the generated output to