LTI-Lib latest version v1.9 - last update 24 Nov 2005
Main Page | Modules | Namespace List | Class Hierarchy | Alphabetical List | Class List | Directories | File List | Namespace Members | Class Members | File Members | Related Pages
lti::clusteringValidity

lti::clusteringValidity Class Reference

Parent class for all clustering validity measures. More...

#include <ltiClusteringValidity.h>

Inheritance diagram for lti::clusteringValidity:

Inheritance graph
[legend]
Collaboration diagram for lti::clusteringValidity:

Collaboration graph
[legend]
List of all members.

Public Member Functions

 clusteringValidity ()
 clusteringValidity (const clusteringValidity &other)
virtual ~clusteringValidity ()
virtual const char * getTypeName () const
clusteringValidity & operator= (const clusteringValidity &other)
clusteringValidity & copy (const clusteringValidity &other)
virtual bool apply (const std::vector< dmatrix > &clusteredData, double &index, const dmatrix &centroids) const =0

Protected Member Functions

double getMinimumDistance (const dmatrix &m1, const dmatrix &m2) const
double getMaximumDistance (const dmatrix &m1, const dmatrix &m2) const
double getCentroidDistance (const dmatrix &m1, const dmatrix &m2) const
double getAverageDistance (const dmatrix &m1, const dmatrix &m2) const
double getAverageInterpointDistance (const dmatrix &m1, const dmatrix &m2) const
double getStandardDiameter (const dmatrix &m1) const
double getAverageDiameter (const dmatrix &m1) const
double getAverageToCentroidDiameter (const dmatrix &m1) const

Detailed Description

Parent class for all clustering validity measures.

Clustering validity measures are used to evaluate the quality of a clustering. The measure can e.a. be used to find the best possible number of clusters in a data set. It provides some distance and diameter measure for clusters. These measures are descriped in IEEE Transaction on systems,man and cybernetics - part B: cybernetics, Vol 28, No.3, June 1998, 301-315

Constructor & Destructor Documentation

lti::clusteringValidity::clusteringValidity (   ) 
 

default constructor

lti::clusteringValidity::clusteringValidity (  const clusteringValidity &  other  ) 
 

default constructor

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

destructor

Member Function Documentation

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

abstract parant class operates on the given parameter.

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
Returns:
true if apply successful or false otherwise.

Implemented in lti::dunnIndex, lti::modHubertStat, lti::normModHubertStat, and lti::daviesBouldinIndex.

clusteringValidity& lti::clusteringValidity::copy (  const clusteringValidity &  other  ) 
 

copy data of "other" functor.

Parameters:
other the functor to be copied
Returns:
a reference to this functor object

double lti::clusteringValidity::getAverageDiameter (  const dmatrix &  m1  )  const [protected]
 

the average distance of all points in m1

double lti::clusteringValidity::getAverageDistance (  const dmatrix &  m1,
const dmatrix &  m2
)  const [protected]
 

return the average distance of all the point in m1 and m2

double lti::clusteringValidity::getAverageInterpointDistance (  const dmatrix &  m1,
const dmatrix &  m2
)  const [protected]
 

accumulates the the distances of all points each matrix to matrix and divides the sum of these distance through the sum of all points

double lti::clusteringValidity::getAverageToCentroidDiameter (  const dmatrix &  m1  )  const [protected]
 

average distance between each point and the centroid of m1

double lti::clusteringValidity::getCentroidDistance (  const dmatrix &  m1,
const dmatrix &  m2
)  const [protected]
 

return the distance of the centroids of m1 and m2

double lti::clusteringValidity::getMaximumDistance (  const dmatrix &  m1,
const dmatrix &  m2
)  const [protected]
 

calculates the maximum distance of all the points in the given matrices and returns this distance

double lti::clusteringValidity::getMinimumDistance (  const dmatrix &  m1,
const dmatrix &  m2
)  const [protected]
 

calculates the minimum distances of the points in the given matrices and returns it.

double lti::clusteringValidity::getStandardDiameter (  const dmatrix &  m1  )  const [protected]
 

return the maximum distance of all points in m1

virtual const char* lti::clusteringValidity::getTypeName (   )  const [virtual]
 

returns the name of this type ("clusteringValidity")

Reimplemented from lti::functor.

Reimplemented in lti::dunnIndex, lti::modHubertStat, lti::normModHubertStat, and lti::daviesBouldinIndex.

clusteringValidity& lti::clusteringValidity::operator= (  const clusteringValidity &  other  ) 
 

alias for copy member

Parameters:
other the functor to be copied
Returns:
a reference to this functor object

The documentation for this class was generated from the following file:
Generated on Thu Nov 24 16:31:05 2005 for LTI-Lib by Doxygen 1.4.4