br.otimizes.oplatool.core.learning

Class SilhouetteIndex

java.lang.Object

br.otimizes.oplatool.core.learning.SilhouetteIndex

All Implemented Interfaces:

java.io.Serializable

public class SilhouetteIndex
extends java.lang.Object
implements java.io.Serializable

It measures the quality of a clustering. Determines how well each object lies within its cluster. A high average silhouette width indicates a good clustering. Computes the average silhouette of observations for different values f k. The Optimal number of clusters k is one that maximize the average silhouette over a range of possible values Kaufman and Rousseauw in 1990 Compute clustering algorithm (e.g., k-means clustering) for different values of k. For instance, by varying k from 1 to 10 clusters. For each k, calculate the average silhouette of observations (avg.sil). Plot the curve of avg.sil according to the number of clusters k. The location of the maximum is considered as the appropriate number of clusters.

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
protected java.util.ArrayList<java.lang.Double>	m_clustersSilhouette Clusters SI.
protected double	m_globalSilhouette Global SI.

Constructor Summary

Constructors

Constructor and Description
SilhouetteIndex() Default constructor.

Method Summary

Modifier and Type	Method and Description
java.lang.String	evalSilhouette(double si) Evaluates a given silhouette index result.
void	evaluate(weka.clusterers.AbstractClusterer clusterer, weka.core.Instances centroids, weka.core.Instances instances, weka.core.DistanceFunction distanceFunction) Evaluates the clusterer quality, using the Silhouette-Index algorithm.
java.util.ArrayList<java.lang.Double>	getClustersSilhouette() Gets the silhouetteIndex for all clusters.
double	getGlobalSilhouette() Gets the global silhouette, i.e: the mean silhouette of all points.
java.lang.String	toString() Returns a string describing the br.otimizes.oplatool.core.jmetal4.results.

Methods inherited from class java.lang.Object

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

Field Detail

m_clustersSilhouette

protected java.util.ArrayList<java.lang.Double> m_clustersSilhouette

Clusters SI.

m_globalSilhouette

protected double m_globalSilhouette

Global SI.

Constructor Detail

SilhouetteIndex

public SilhouetteIndex()

Default constructor.

Method Detail

evaluate

public void evaluate(weka.clusterers.AbstractClusterer clusterer,
                     weka.core.Instances centroids,
                     weka.core.Instances instances,
                     weka.core.DistanceFunction distanceFunction)
              throws java.lang.Exception

Evaluates the clusterer quality, using the Silhouette-Index algorithm.

Parameters:

clusterer - given clusterer.

instances - dataset.

Throws:

java.lang.Exception

getClustersSilhouette

public java.util.ArrayList<java.lang.Double> getClustersSilhouette()

Gets the silhouetteIndex for all clusters.

Returns:

Returns the clusters Silhouette-Index.

getGlobalSilhouette

public double getGlobalSilhouette()

Gets the global silhouette, i.e: the mean silhouette of all points.

Returns:

Returns the mean silhouette of all points.

evalSilhouette

public java.lang.String evalSilhouette(double si)

Evaluates a given silhouette index result.

Parameters:

si - Silhouette-Index.

toString

public java.lang.String toString()

Returns a string describing the br.otimizes.oplatool.core.jmetal4.results.

Overrides:

toString in class java.lang.Object

Returns:

a string describing the clusterer.