lti::principalComponents< T >::parameters Class Reference

the parameters for the class principalComponents More...

#include <ltiPCA.h>

Inheritance diagram for lti::principalComponents< T >::parameters:

[legend]

Collaboration diagram for lti::principalComponents< T >::parameters:

[legend]

List of all members.

Public Member Functions
	parameters ()
	parameters (const parameters &other)
	~parameters ()
const char *	getTypeName () const
parameters &	copy (const parameters &other)
parameters &	operator= (const parameters &other)
virtual functor::parameters *	clone () const
virtual bool	read (ioHandler &handler, const bool complete=true)
virtual bool	write (ioHandler &handler, const bool complete=true) const
Public Attributes
int	resultDim
bool	autoDim
bool	useCorrelation
bool	whitening
eigenSystem< T > *	eigen
T	relevance
bool	centerData
bool	useSVD
singularValueDecomp< T > *	svd

---

Detailed Description

template<class T>
class lti::principalComponents< T >::parameters

the parameters for the class principalComponents

---

Constructor & Destructor Documentation

template<class T>

lti::principalComponents< T >::parameters::parameters

(

void

)

[inline]

default constructor

Reimplemented from lti::linearAlgebraFunctor::parameters.

Reimplemented in lti::regularizedPCA< T >::parameters.

References lti::principalComponents< T >::parameters::autoDim, lti::principalComponents< T >::parameters::centerData, lti::principalComponents< T >::parameters::eigen, lti::principalComponents< T >::parameters::relevance, lti::principalComponents< T >::parameters::resultDim, lti::principalComponents< T >::parameters::svd, lti::principalComponents< T >::parameters::useCorrelation, lti::principalComponents< T >::parameters::useSVD, and lti::principalComponents< T >::parameters::whitening.

Referenced by lti::principalComponents< T >::parameters::clone().

template<class T>

lti::principalComponents< T >::parameters::parameters

(

const parameters &

other

)

[inline]

copy constructor

Parameters:

other

the parameters object to be copied

Reimplemented from lti::linearAlgebraFunctor::parameters.

Reimplemented in lti::regularizedPCA< T >::parameters.

References lti::principalComponents< T >::parameters::copy(), lti::principalComponents< T >::parameters::eigen, and lti::principalComponents< T >::parameters::svd.

template<class T>

lti::principalComponents< T >::parameters::~parameters

(

)

[inline, virtual]

destructor

Reimplemented from lti::linearAlgebraFunctor::parameters.

Reimplemented in lti::regularizedPCA< T >::parameters.

References lti::principalComponents< T >::parameters::eigen.

---

Member Function Documentation

template<class T>

virtual functor::parameters* lti::principalComponents< T >::parameters::clone

(

)

const [inline, virtual]

returns a pointer to a clone of the parameters

Reimplemented from lti::linearAlgebraFunctor::parameters.

Reimplemented in lti::regularizedPCA< T >::parameters.

References lti::principalComponents< T >::parameters::parameters().

template<class T>

parameters& lti::principalComponents< T >::parameters::copy

(

const parameters &

other

)

[inline]

copy the contents of a parameters object

Parameters:

other

the parameters object to be copied

Returns:

a reference to this parameters object

Reimplemented from lti::linearAlgebraFunctor::parameters.

Reimplemented in lti::regularizedPCA< T >::parameters.

References lti::principalComponents< T >::parameters::autoDim, lti::principalComponents< T >::parameters::centerData, lti::functor::parameters::clone(), lti::functor::parameters::copy(), lti::principalComponents< T >::parameters::eigen, lti::principalComponents< T >::parameters::relevance, lti::principalComponents< T >::parameters::resultDim, lti::principalComponents< T >::parameters::svd, lti::principalComponents< T >::parameters::useCorrelation, lti::principalComponents< T >::parameters::useSVD, and lti::principalComponents< T >::parameters::whitening.

Referenced by lti::principalComponents< T >::parameters::operator=(), and lti::principalComponents< T >::parameters::parameters().

template<class T>

const char* lti::principalComponents< T >::parameters::getTypeName

(

void

)

const [inline, virtual]

returns name of this type

Reimplemented from lti::linearAlgebraFunctor::parameters.

Reimplemented in lti::regularizedPCA< T >::parameters.

template<class T>

parameters& lti::principalComponents< T >::parameters::operator=

(

const parameters &

other

)

[inline]

Assigns the contents of the other object to this object.

Reimplemented in lti::regularizedPCA< T >::parameters.

References lti::principalComponents< T >::parameters::copy().

template<class T>

virtual bool lti::principalComponents< T >::parameters::read	(	ioHandler &	handler,
		const bool	complete = true
	)			[inline, virtual]

read the parameters from the given ioHandler

Parameters:

	handler	the ioHandler to be used
	complete	if true (the default) the enclosing begin/end will be also read, otherwise only the data block will be read.

Returns:

true if write was successful

Reimplemented from lti::linearAlgebraFunctor::parameters.

Reimplemented in lti::regularizedPCA< T >::parameters.

References lti::principalComponents< T >::parameters::autoDim, lti::principalComponents< T >::parameters::centerData, lti::ioHandler::readBegin(), lti::ioHandler::readEnd(), lti::principalComponents< T >::parameters::relevance, lti::principalComponents< T >::parameters::resultDim, lti::principalComponents< T >::parameters::useCorrelation, and lti::principalComponents< T >::parameters::whitening.

template<class T>

virtual bool lti::principalComponents< T >::parameters::write	(	ioHandler &	handler,
		const bool	complete = true
	)			const [inline, virtual]

write the parameters in the given ioHandler

Parameters:

	handler	the ioHandler to be used
	complete	if true (the default) the enclosing begin/end will be also written, otherwise only the data block will be written.

Returns:

true if write was successful

Reimplemented from lti::linearAlgebraFunctor::parameters.

Reimplemented in lti::regularizedPCA< T >::parameters.

References lti::principalComponents< T >::parameters::autoDim, lti::principalComponents< T >::parameters::centerData, lti::principalComponents< T >::parameters::relevance, lti::principalComponents< T >::parameters::resultDim, lti::principalComponents< T >::parameters::useCorrelation, lti::principalComponents< T >::parameters::whitening, lti::ioHandler::writeBegin(), and lti::ioHandler::writeEnd().

---

Member Data Documentation

template<class T>

bool lti::principalComponents< T >::parameters::autoDim

This flag determines, if the functor should determine a maximum allowable dimension itself.

"Maximum dimension" means that the dimension is equal to the number of eigenvalues of the covariance matrix which are larger than zero.

Default value: false

Referenced by lti::principalComponents< T >::parameters::copy(), lti::principalComponents< T >::parameters::parameters(), lti::principalComponents< T >::parameters::read(), and lti::principalComponents< T >::parameters::write().

template<class T>

bool lti::principalComponents< T >::parameters::centerData

This flag denotes if the transformed data should be centered around zero.

This is the usual behaviour of the PCA, but for some special operations it may be necessary to disable this. If the flag is false, the mean of the transformed data is moved to the transformed mean of the source data.

Default: true

Referenced by lti::principalComponents< T >::parameters::copy(), lti::principalComponents< T >::parameters::parameters(), lti::principalComponents< T >::parameters::read(), and lti::principalComponents< T >::parameters::write().

template<class T>

eigenSystem<T>* lti::principalComponents< T >::parameters::eigen

The eigensystem functor used for computing eigenvectors.

Default with LAPACK: lti::fastEigenSystem<T> Default without LAPACK: lti::jacobi<T>

Referenced by lti::principalComponents< T >::parameters::copy(), lti::principalComponents< T >::parameters::parameters(), and lti::principalComponents< T >::parameters::~parameters().

template<class T>

T lti::principalComponents< T >::parameters::relevance

The factor which determines relevant eigenvectors.

An eigenvector is considered relevant if its eigenvalue is at least as large as the largest eigenvalue divided by this number. Usually, it takes values between 1e4 and 1e6.

Default: T(100000)

Referenced by lti::principalComponents< T >::parameters::copy(), lti::principalComponents< T >::parameters::parameters(), lti::principalComponents< T >::parameters::read(), and lti::principalComponents< T >::parameters::write().

template<class T>

int lti::principalComponents< T >::parameters::resultDim

This is the dimension of the reduced vectors.

Default value: 3

Referenced by lti::principalComponents< T >::parameters::copy(), lti::principalComponents< T >::parameters::parameters(), lti::principalComponents< T >::parameters::read(), and lti::principalComponents< T >::parameters::write().

template<class T>

singularValueDecomp<T>* lti::principalComponents< T >::parameters::svd

Singular value decomposition to be used when useSVD is true.

Default is null. If LAPACK is used and data matrices are large consider using lti::fastSVD<T>.

Referenced by lti::principalComponents< T >::parameters::copy(), and lti::principalComponents< T >::parameters::parameters().

template<class T>

bool lti::principalComponents< T >::parameters::useCorrelation

This flag determines if the functor should use the correlation coefficient matrix (flag is true) for eigenvector computation or the covariance matrix (flag is false).

The default is false.

Referenced by lti::principalComponents< T >::parameters::copy(), lti::principalComponents< T >::parameters::parameters(), lti::principalComponents< T >::parameters::read(), and lti::principalComponents< T >::parameters::write().

template<class T>

bool lti::principalComponents< T >::parameters::useSVD

When true singular value decomposition instead of eigensystem solution is used to calculate the eigenvectors and eigenvalues.

This can be much faster and less memory consuming. Default is false.

Referenced by lti::principalComponents< T >::parameters::copy(), and lti::principalComponents< T >::parameters::parameters().

template<class T>

bool lti::principalComponents< T >::parameters::whitening

This flag determines if the functor should perform a whitening transform of the data.

Whitening means that 1. A PCA is performed, using the covariance matrix for eigenvector computation 2. A scaling of the transformed data by the inverse of the square root of the eigenvalues.

You should set useCorrelation to false if you use whitening.

The default is false.

Referenced by lti::principalComponents< T >::parameters::copy(), lti::principalComponents< T >::parameters::parameters(), lti::principalComponents< T >::parameters::read(), and lti::principalComponents< T >::parameters::write().

---

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

• ltiPCA.h