lti::mahalanobisDistOfSubset Class Reference

This class implements a cost function for evaluating a subset of features based on the mahalanobis distance. More...

#include <ltiMahalanobisDistOfSubset.h>

Inheritance diagram for lti::mahalanobisDistOfSubset:

[legend]

Collaboration diagram for lti::mahalanobisDistOfSubset:

[legend]

List of all members.

Public Member Functions
	mahalanobisDistOfSubset ()
	mahalanobisDistOfSubset (const mahalanobisDistOfSubset &other)
virtual	~mahalanobisDistOfSubset ()
virtual const char *	getTypeName () const
virtual void	setSrc (const dmatrix &src, const ivector &srcIds)
virtual bool	apply (const std::list< int > &in, double &value)
mahalanobisDistOfSubset &	copy (const mahalanobisDistOfSubset &other)
mahalanobisDistOfSubset &	operator= (const mahalanobisDistOfSubset &other)
virtual functor *	clone () const
const parameters &	getParameters () const
Protected Attributes
list< dmatrix >	clusters

---

Detailed Description

This class implements a cost function for evaluating a subset of features based on the mahalanobis distance.

A great value of this cost function shows well seperated clusters. First, for all points in the data set the mahalanobis distance to each cluster in distribution is computed. Then the result of this costFunction is the average of all these distances. Before you can compute the cost function use the setSrc method to pass the complete data set with all features to this class. Later the features are accessed only by their column number.

---

Constructor & Destructor Documentation

lti::mahalanobisDistOfSubset::mahalanobisDistOfSubset

(

)

default constructor

lti::mahalanobisDistOfSubset::mahalanobisDistOfSubset

(

const mahalanobisDistOfSubset &

other

)

copy constructor

Parameters:

other

the object to be copied

virtual lti::mahalanobisDistOfSubset::~mahalanobisDistOfSubset

(

)

[virtual]

destructor

---

Member Function Documentation

virtual bool lti::mahalanobisDistOfSubset::apply	(	const std::list< int > &	in,
		double &	value
	)			[virtual]

computes the mahalanobis distance with the given subset of features

Parameters:

	value	the result of the cost function.
	in	std::list with the features that should be inclueded for computing the function

Returns:

true if apply successful or false otherwise.

Implements lti::costFunction.

virtual functor* lti::mahalanobisDistOfSubset::clone

(

)

const [virtual]

returns a pointer to a clone of this functor.

Implements lti::functor.

mahalanobisDistOfSubset& lti::mahalanobisDistOfSubset::copy

(

const mahalanobisDistOfSubset &

other

)

copy data of "other" functor.

Parameters:

other

the functor to be copied

Returns:

a reference to this functor object

Reimplemented from lti::costFunction.

const parameters& lti::mahalanobisDistOfSubset::getParameters

(

)

const

returns used parameters

Reimplemented from lti::functor.

virtual const char* lti::mahalanobisDistOfSubset::getTypeName

(

)

const [virtual]

returns the name of this type ("mahalanobisDistOfSubset")

Reimplemented from lti::costFunction.

mahalanobisDistOfSubset& lti::mahalanobisDistOfSubset::operator=

(

const mahalanobisDistOfSubset &

other

)

alias for copy member

Parameters:

other

the functor to be copied

Returns:

a reference to this functor object

Reimplemented from lti::costFunction.

virtual void lti::mahalanobisDistOfSubset::setSrc	(	const dmatrix &	src,
		const ivector &	srcIds
	)			[virtual]

Set the src data, that is used for computation.

This must be done before apply methods are called. Later each feature can be adressed by its column number, because usually a column represents a feature.

Reimplemented from lti::costFunction.

---

Member Data Documentation

list<dmatrix> lti::mahalanobisDistOfSubset::clusters [protected]

A list with each element containing a matrix with one cluster.

This list is build from the contructor.

---

The documentation for this class was generated from the following file:

• ltiMahalanobisDistOfSubset.h