idl.tmt.clusterers
Class BestKMeansEMInitializer

java.lang.Object
  |
  +--idl.tmt.clusterers.BestKMeansEMInitializer

All Implemented Interfaces:

EMInitializer, java.io.Serializable

public class BestKMeansEMInitializer

extends java.lang.Object

implements EMInitializer

EMInitializer class which runs K-means ten times and initializes the EM model to the cluster centroids of the best K-means result. This initialization does NOT support datasets with nominal attributes. Created on Jun 10, 2004

Author:

jelsas

See Also:

Serialized Form

Field Summary

private int

numIterations

Constructor Summary

BestKMeansEMInitializer()

Method Summary

int	getNumIterations()
void	initializeModel(int numClusters, double minStdDev, Instances data, java.util.Random random, double[] attributeMinVals, double[] attributeMaxVals, double[][][] modelNormal, double[][] weights, double[] priors)
void	setNumIterations(int numIterations)

Methods inherited from class java.lang.Object

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

Field Detail

numIterations

private int numIterations

Constructor Detail

BestKMeansEMInitializer

public BestKMeansEMInitializer()

Method Detail

initializeModel

public void initializeModel(int numClusters,
                            double minStdDev,
                            Instances data,
                            java.util.Random random,
                            double[] attributeMinVals,
                            double[] attributeMaxVals,
                            double[][][] modelNormal,
                            double[][] weights,
                            double[] priors)
                     throws java.lang.Exception

Specified by:

initializeModel in interface EMInitializer

setNumIterations

public void setNumIterations(int numIterations)

getNumIterations

public int getNumIterations()