lti::normModHubertStat Class Reference

Calculates the normalized version of the modified Hubert statistic The index is between -1 and 1. More...

#include <ltiClusteringValidity.h>

Inheritance diagram for lti::normModHubertStat:

[legend]

Collaboration diagram for lti::normModHubertStat:

[legend]

List of all members.

Public Member Functions
	normModHubertStat ()
	normModHubertStat (const normModHubertStat &other)
virtual	~normModHubertStat ()
normModHubertStat &	copy (const normModHubertStat &other)
normModHubertStat &	operator= (const normModHubertStat &other)
virtual const char *	getTypeName () const
virtual functor *	clone () const
virtual bool	apply (const std::vector< dmatrix > &clusteredData, double &index, const dmatrix &centroids) const

Detailed Description

Calculates the normalized version of the modified Hubert statistic The index is between -1 and 1.

A value near 1 shows compact well separated clusters. When the same data set is tested with different parameters, e.a. with different number of clusters, the best possible clustering is shown by a sharp knee in this index. This validity measure is not defined, if the number of cluster is 1 or equal to the number of points

Constructor & Destructor Documentation

lti::normModHubertStat::normModHubertStat ( )

standard contructor

lti::normModHubertStat::normModHubertStat ( const normModHubertStat & other )

standard contructor

virtual lti::normModHubertStat::~normModHubertStat ( ) [virtual]

destructor

Member Function Documentation

virtual bool lti::normModHubertStat::apply	(	const std::vector< dmatrix > &	clusteredData,
		double &	index,
		const dmatrix &	centroids
	)			const `[virtual]`

calculates the normalized modified Hubert Statistics of the clustered data.

Parameters:

	clusteredData	std::vector<dmatrix> with the source data.
	index	the clustering validity result
	centroids	dmatrix with each row representing a centroid of the corresponding distribution in clusteredData