Multivariate gaussian mixture. More...

#include <CRNMultivariateGaussianMixture.h>

Inheritance diagram for crn::MultivariateGaussianMixture:

Collaboration diagram for crn::MultivariateGaussianMixture:

Public Member Functions
	MultivariateGaussianMixture ()
	Default constructor. More...

	MultivariateGaussianMixture (size_t d)
	Default constructor. More...

	MultivariateGaussianMixture (const MultivariateGaussianMixture &)=default

	MultivariateGaussianMixture (MultivariateGaussianMixture &&)=default

template<typename ITER >
	MultivariateGaussianMixture (ITER it_begin, ITER it_end, size_t nb_seeds=2)

MultivariateGaussianMixture &	operator= (const MultivariateGaussianMixture &)=default

MultivariateGaussianMixture &	operator= (MultivariateGaussianMixture &&)=default

virtual	~MultivariateGaussianMixture () override
	Destructor. More...

void	SetTo (const MultivariateGaussianMixture &m)
	Set mixture from another one. More...

size_t	GetDimension () const noexcept
	Returns the number of features. More...

void	SetDimension (size_t k) noexcept
	Sets the number of features. More...

size_t	GetNbMembers () const noexcept
	Returns the number of density functions. More...

double	GetWeight (size_t k) const
	Returns the weight of a given density function. More...

MultivariateGaussianPDF	GetMember (size_t k) const
	Returns a given density function. More...

MatrixDouble	GetMean (size_t k) const
	Returns the mean of a given density function. More...

SquareMatrixDouble	GetVariance (size_t k) const
	Returns the variance of a given density function. More...

void	AddMember (const MultivariateGaussianPDF &pdf, double Weight)
	Adds a density function. More...

void	SetMember (const MultivariateGaussianPDF &pdf, double w, size_t k)
	Replaces a given density function and its weight. More...

double	ValueAt (const MatrixDouble &X) const
	Evaluates a pattern. More...

double	ValueAt (const MatrixDouble &X, size_t k, bool w=false) const
	Evaluates a pattern for a given density function. More...

double	ValueAt (const std::vector< double > &x, size_t k, bool w=false) const
	Evaluates a pattern for a given density function. More...

unsigned int	EM (const MatrixDouble &patterns, size_t nbSeeds=2, double epsilon=std::numeric_limits< double >::epsilon(), size_t MaximalIterations=100)
	Expectation Maximization. More...

unsigned int	EM (const std::vector< std::vector< double > > &patterns, size_t nbSeeds=2, double epsilon=std::numeric_limits< double >::epsilon(), size_t MaximalIterations=100)

template<typename ITER >
unsigned int	EM (ITER it_begin, ITER it_end, size_t nbSeeds=2, double epsilon=std::numeric_limits< double >::epsilon(), size_t MaximalIterations=100)

String	ToString () const
	Dumps a summary of the mixture to a string. More...

String	ToString (size_t k) const
	Dumps a summary of one element of the mixture to a string. More...

double	MLLE (const MatrixDouble &data) const
	Maximum log-likelihood estimation. More...

double	MLLE (const std::vector< std::vector< double > > &data) const
	Maximum log-likelihood estimation. More...

template<typename ITER >
double	MLLE (ITER it_begin, ITER it_end) const
	Maximum log-likelihood estimation. More...

bool	IsValid () const
	Test if mixture is valid i.e. if all paratemers have finite values. More...

Public Member Functions inherited from crn::Object
virtual	~Object ()=default

Detailed Description

Multivariate gaussian mixture.

Model for multivariate gaussian mixture

Author: Jean DUONG

Date: August 2008

Version: 0.2

Definition at line 43 of file CRNMultivariateGaussianMixture.h.

Constructor & Destructor Documentation

crn::MultivariateGaussianMixture::MultivariateGaussianMixture ( )

inline

Default constructor.

Definition at line 47 of file CRNMultivariateGaussianMixture.h.

crn::MultivariateGaussianMixture::MultivariateGaussianMixture ( size_t d )

inline

Default constructor.

Definition at line 49 of file CRNMultivariateGaussianMixture.h.

crn::MultivariateGaussianMixture::MultivariateGaussianMixture ( const MultivariateGaussianMixture & )

default

crn::MultivariateGaussianMixture::MultivariateGaussianMixture ( MultivariateGaussianMixture && )

default

template<typename ITER >

crn::MultivariateGaussianMixture::MultivariateGaussianMixture	(	ITER	it_begin,
		ITER	it_end,
		size_t	nb_seeds = `2`
	)

inline

Create a gaussian mixture to model a given set of data

Parameters

[in]	it_begin	iterator pointing to first (vector of double, size_t) pair of input data
[in]	it_end	ending iterator
[in]	nbSeeds	the number of density functions

Returns: the inferred gaussian mixture model

Definition at line 62 of file CRNMultivariateGaussianMixture.h.

MultivariateGaussianMixture::~MultivariateGaussianMixture ( )

overridevirtual

Destructor.

Destructor

Definition at line 41 of file CRNMultivariateGaussianMixture.cpp.

Member Function Documentation

void MultivariateGaussianMixture::AddMember	(	const MultivariateGaussianPDF &	pdf,
		double	w
	)

Adds a density function.

Add a member in mixture

Exceptions

ExceptionDimension incompatible dimensions

Parameters

[in]	spdf	gaussian PDF to add as new member
[in]	w	Weight of the new member

Definition at line 155 of file CRNMultivariateGaussianMixture.cpp.

unsigned int MultivariateGaussianMixture::EM	(	const MatrixDouble &	patterns,
		size_t	nbSeeds = `2`,
		double	epsilon = `std::numeric_limits<double>::epsilon()`,
		size_t	maximalIterations = `100`
	)

Expectation Maximization.

Expectation-Maximization algorithm for Gaussian mixture estimation optimization

Parameters

[in]	patterns	set of patterns stored as rows of a data matrix
[in]	nbSeeds	the number of density functions
[in]	epsilon	the minimum likelihood gain between two iterations
[in]	maximalIterations	the maximum number of iterations

Returns: the number of iterations done to optimize the mixture

Definition at line 324 of file CRNMultivariateGaussianMixture.cpp.

unsigned int MultivariateGaussianMixture::EM	(	const std::vector< std::vector< double > > &	patterns,
		size_t	nbSeeds = `2`,
		double	epsilon = `std::numeric_limits<double>::epsilon()`,
		size_t	maximalIterations = `100`
	)

Expectation-Maximization algorithm for Gaussian mixture estimation optimization

Parameters

[in]	patterns	vector of patterns stored as vectors of doubles
[in]	nbSeeds	the number of density functions
[in]	epsilon	the minimum likelihood gain between two iterations
[in]	maximalIterations	the maximum number of iterations

Returns: the number of iterations done to optimize the mixture

Definition at line 495 of file CRNMultivariateGaussianMixture.cpp.

template<typename ITER >

unsigned int crn::MultivariateGaussianMixture::EM	(	ITER	it_begin,
		ITER	it_end,
		size_t	nbSeeds = `2`,
		double	epsilon = `std::numeric_limits<double>::epsilon()`,
		size_t	MaximalIterations = `100`
	)

inline

Expectation-Maximization algorithm for Gaussian mixture estimation optimization

Parameters

[in]	it_begin	iterator pointing to first (vector of double, size_t) pair of input data
[in]	it_end	ending iterator
[in]	nbSeeds	the number of density functions
[in]	epsilon	the minimum likelihood gain between two iterations
[in]	maximalIterations	the maximum number of iterations

Returns: the number of iterations done to optimize the mixture

Definition at line 122 of file CRNMultivariateGaussianMixture.h.

size_t crn::MultivariateGaussianMixture::GetDimension ( ) const

inlinenoexcept

Returns the number of features.

Definition at line 79 of file CRNMultivariateGaussianMixture.h.

MatrixDouble MultivariateGaussianMixture::GetMean ( size_t k ) const

Returns the mean of a given density function.

Get mean for a member

Exceptions

ExceptionDomain index out of bounds

Parameters

[in] k index of a density function %

Returns: the mean of the density function

Definition at line 121 of file CRNMultivariateGaussianMixture.cpp.

MultivariateGaussianPDF MultivariateGaussianMixture::GetMember ( size_t k ) const

Returns a given density function.

Return a member

Exceptions

ExceptionDomain index out of bounds

Parameters

[in]

k

index of the density function to return

Returns: the kth density function

Definition at line 85 of file CRNMultivariateGaussianMixture.cpp.

size_t crn::MultivariateGaussianMixture::GetNbMembers ( ) const

inlinenoexcept

Returns the number of density functions.

Definition at line 83 of file CRNMultivariateGaussianMixture.h.

SquareMatrixDouble MultivariateGaussianMixture::GetVariance ( size_t k ) const

Returns the variance of a given density function.

Get variance for a member

Exceptions

ExceptionDomain index out of bounds

Parameters

[in] k index of a density function

Returns: the variance of the density function

Definition at line 139 of file CRNMultivariateGaussianMixture.cpp.

double MultivariateGaussianMixture::GetWeight ( size_t k ) const

Returns the weight of a given density function.

Get weight for a member

Exceptions

ExceptionDomain index out of bounds

Parameters

[in] k index of a density function %

Returns: the weight of the density function

Definition at line 103 of file CRNMultivariateGaussianMixture.cpp.

bool MultivariateGaussianMixture::IsValid ( ) const

Test if mixture is valid i.e. if all paratemers have finite values.

Test if mixture is valid i.e. if all paratemers have finite values

Returns: true if success, false else

Definition at line 708 of file CRNMultivariateGaussianMixture.cpp.

double MultivariateGaussianMixture::MLLE ( const MatrixDouble & Data ) const

Maximum log-likelihood estimation.

Maximum log-likelihood estimator

Parameters

[in] Data set of patterns stored as rows of a data matrix

Returns: the likelihood that the mixture modelizes the set of patterns

Definition at line 279 of file CRNMultivariateGaussianMixture.cpp.

double MultivariateGaussianMixture::MLLE ( const std::vector< std::vector< double > > & data ) const

Maximum log-likelihood estimation.

Maximum log-likelihood estimator

Parameters

[in] Data set of patterns stored as rows of a data matrix

Returns: the likelihood that the mixture modelizes the set of patterns

Definition at line 304 of file CRNMultivariateGaussianMixture.cpp.

template<typename ITER >

double crn::MultivariateGaussianMixture::MLLE	(	ITER	it_begin,
		ITER	it_end
	)		const

inline

Maximum log-likelihood estimation.

Definition at line 319 of file CRNMultivariateGaussianMixture.h.

MultivariateGaussianMixture& crn::MultivariateGaussianMixture::operator= ( const MultivariateGaussianMixture & )

default

MultivariateGaussianMixture& crn::MultivariateGaussianMixture::operator= ( MultivariateGaussianMixture && )

default

void MultivariateGaussianMixture::SetDimension ( size_t k )

noexcept

Sets the number of features.

Set dimension

Parameters

[in] k dimension of data space

Definition at line 72 of file CRNMultivariateGaussianMixture.cpp.

void MultivariateGaussianMixture::SetMember	(	const MultivariateGaussianPDF &	pdf,
		double	w,
		size_t	k
	)

Replaces a given density function and its weight.

Replace a member in mixture

Exceptions

ExceptionDimension	incompatible dimensions
ExceptionDomain	invalid index

Parameters

[in]	spdf	the new gaussian PDF
[in]	w	the new weight
[in]	k	index

Definition at line 177 of file CRNMultivariateGaussianMixture.cpp.

void MultivariateGaussianMixture::SetTo ( const MultivariateGaussianMixture & m )

Set mixture from another one.

Set up parameters from another mixture's

Parameters

[in] *m another mixture

Definition at line 51 of file CRNMultivariateGaussianMixture.cpp.

String MultivariateGaussianMixture::ToString ( ) const

Dumps a summary of the mixture to a string.

Dump the mixture parameters to a string

Returns: the string

Definition at line 665 of file CRNMultivariateGaussianMixture.cpp.

String MultivariateGaussianMixture::ToString ( size_t k ) const

Dumps a summary of one element of the mixture to a string.

Dump the parameters of one member in the mixture to a string

Parameters

[in] k the index of the density function

Returns: the string

Definition at line 682 of file CRNMultivariateGaussianMixture.cpp.

double MultivariateGaussianMixture::ValueAt ( const MatrixDouble & X ) const

Evaluates a pattern.

Evaluate a pattern

Parameters

[in] X the column matrix to evaluate

Returns: the image of X

Definition at line 200 of file CRNMultivariateGaussianMixture.cpp.

double MultivariateGaussianMixture::ValueAt	(	const MatrixDouble &	X,
		size_t	k,
		bool	w = `false`
	)		const

Evaluates a pattern for a given density function.

Evaluate a pattern only for one member of the mixture

Exceptions

ExceptionDomain index out of bounds

Parameters

[in]	X	the column matrix to evaluate
[in]	k	member index
[in]	w	flag to indicate if weight is used

Returns: the image of X though the kth density function

Definition at line 221 of file CRNMultivariateGaussianMixture.cpp.

double MultivariateGaussianMixture::ValueAt	(	const std::vector< double > &	x,
		size_t	k,
		bool	w = `false`
	)		const

Evaluates a pattern for a given density function.

Evaluate a pattern only for one member of the mixture

Exceptions

ExceptionDomain index out of bounds

Parameters

[in]	x	pattern vector to evaluate
[in]	k	member index
[in]	w	flag to indicate if weight is used

Returns: the image of X though the kth density function

Definition at line 251 of file CRNMultivariateGaussianMixture.cpp.

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

/home/travis/build/Liris-Pleiad/libcrn/src/libcrn/CRNMath/CRNMultivariateGaussianMixture.h
/home/travis/build/Liris-Pleiad/libcrn/src/libcrn/CRNMath/CRNMultivariateGaussianMixture.cpp

Public Member Functions

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation