Discrete HMM class. More...

#include <CRNDiscreteHMM.h>

Inheritance diagram for crn::DiscreteHMM:

Collaboration diagram for crn::DiscreteHMM:

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

	DiscreteHMM (int nstates, int nsymbs)
	Constructor. More...

	DiscreteHMM (const DiscreteHMM &h)
	Copy constructor. More...

DiscreteHMM &	operator= (const DiscreteHMM &)=delete

	DiscreteHMM (DiscreteHMM &&) noexcept

DiscreteHMM &	operator= (DiscreteHMM &&) noexcept

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

bool	operator== (const DiscreteHMM &other) const
	Tests the equality of two models. More...

size_t	GetNbStates () const noexcept
	Returns the number of states. More...

size_t	GetNbSymbols () const noexcept
	Returns the number of symbols. More...

SCSquareMatrixDouble	GetStateTransitionProbability () const
	Returns a reference to the state transition probability matrix. More...

SCMatrixDouble	GetStateGivenSymbolProbability () const
	Returns a reference to the state given symbol probability matrix. More...

SCMatrixDouble	GetFirstStateProbability () const
	Returns a reference to the first state probability matrix. More...

void	SetStateTransitionProbability (const SquareMatrixDouble &a)
	Sets the state transition probability matrix. More...

void	SetStateGivenSymbolProbability (const MatrixDouble &b)
	Sets the state given symbol probability matrix. More...

void	SetFirstStateProbability (const MatrixDouble &p)
	Sets the first state probability matrix. More...

bool	IsValid () const
	Checks if the model is valid. More...

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

double	SequenceProbability (const MatrixInt &observed) const
	Returns the a priori probability of an observed sequence. More...

UMatrixInt	MakeViterbi (const MatrixInt &observed) const
	Returns the most likely state state sequence corresponding to an observed sequence. More...

void	BaumWelchSingle (const MatrixInt &observed, size_t maxIter)
	Learning from one observed sequence. More...

void	BaumWelchMultiple (const MatrixInt &observationSet, size_t maxIter)
	Learning from multiple observed sequences. More...

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

Detailed Description

Discrete HMM class.

Discrete Hidden Markov Model

Author: Jean DUONG

Date: September 2008

Version: 0.3

Definition at line 41 of file CRNDiscreteHMM.h.

Constructor & Destructor Documentation

crn::DiscreteHMM::DiscreteHMM ( )

inline

Default constructor.

Definition at line 46 of file CRNDiscreteHMM.h.

DiscreteHMM::DiscreteHMM	(	int	nstates,
		int	nsymbs
	)

Constructor.

Constructor

Parameters

[in]	nstates	number of states for HMM
[in]	nsymbs	number of symbols for HMM

Definition at line 56 of file CRNDiscreteHMM.cpp.

DiscreteHMM::DiscreteHMM ( const DiscreteHMM & h )

Copy constructor.

Copy constructor

Parameters

[in] h another HMM

Definition at line 40 of file CRNDiscreteHMM.cpp.

DiscreteHMM::DiscreteHMM ( DiscreteHMM && )

defaultnoexcept

DiscreteHMM::~DiscreteHMM ( )

overridevirtual

Destructor.

Destructor

Definition at line 71 of file CRNDiscreteHMM.cpp.

Member Function Documentation

void DiscreteHMM::BaumWelchMultiple	(	const MatrixInt &	observationSet,
		size_t	maxIter
	)

Learning from multiple observed sequences.

HMM model training with Baum-Welch procedure. Training with many observation sequences.

Parameters

[in]	observationSet	the observed sequences. Each sequence is a line of the matrix.
[in]	maxIter	maximal number of iterations

Definition at line 507 of file CRNDiscreteHMM.cpp.

void DiscreteHMM::BaumWelchSingle	(	const MatrixInt &	observed,
		size_t	maxIter
	)

Learning from one observed sequence.

HMM model training with Baum-Welch procedure. Training with one observation sequence.

Parameters

[in]	observed	the observed sequence as a line matrix
[in]	maxIter	the maximum number of iterations

Definition at line 350 of file CRNDiscreteHMM.cpp.

SCMatrixDouble crn::DiscreteHMM::GetFirstStateProbability ( ) const

inline

Returns a reference to the first state probability matrix.

Definition at line 73 of file CRNDiscreteHMM.h.

size_t crn::DiscreteHMM::GetNbStates ( ) const

inlinenoexcept

Returns the number of states.

Definition at line 63 of file CRNDiscreteHMM.h.

size_t crn::DiscreteHMM::GetNbSymbols ( ) const

inlinenoexcept

Returns the number of symbols.

Definition at line 66 of file CRNDiscreteHMM.h.

SCMatrixDouble crn::DiscreteHMM::GetStateGivenSymbolProbability ( ) const

inline

Returns a reference to the state given symbol probability matrix.

Definition at line 71 of file CRNDiscreteHMM.h.

SCSquareMatrixDouble crn::DiscreteHMM::GetStateTransitionProbability ( ) const

inline

Returns a reference to the state transition probability matrix.

Definition at line 69 of file CRNDiscreteHMM.h.

bool DiscreteHMM::IsValid ( ) const

Checks if the model is valid.

Check if model is valid

Returns: true if the model is valid, false else

Definition at line 114 of file CRNDiscreteHMM.cpp.

UMatrixInt DiscreteHMM::MakeViterbi ( const MatrixInt & observed ) const

Returns the most likely state state sequence corresponding to an observed sequence.

Viterbi procedure to estimate the best state chain given an observation chain

Parameters

[in] observed the observed sequence

Returns: the most likely state sequence corresponding to the observed sequence

Definition at line 270 of file CRNDiscreteHMM.cpp.

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

delete

DiscreteHMM & DiscreteHMM::operator= ( DiscreteHMM && )

defaultnoexcept

bool DiscreteHMM::operator== ( const DiscreteHMM & other ) const

Tests the equality of two models.

Check if two models are equal

Parameters

[in] other the other HMM

Returns: true if models are equal, false else

Definition at line 141 of file CRNDiscreteHMM.cpp.

double DiscreteHMM::SequenceProbability ( const MatrixInt & observed ) const

Returns the a priori probability of an observed sequence.

Probability estimation for a given observation sequence

Parameters

[in] observed the observed sequence

Returns: the a priori probability of the observed sequence

Definition at line 247 of file CRNDiscreteHMM.cpp.

void DiscreteHMM::SetFirstStateProbability ( const MatrixDouble & p )

Sets the first state probability matrix.

Set the first state probability matrix. A copy of the matrix is made.

Parameters

[in] p the new first state probability matrix

Definition at line 103 of file CRNDiscreteHMM.cpp.

void DiscreteHMM::SetStateGivenSymbolProbability ( const MatrixDouble & b )

Sets the state given symbol probability matrix.

Set the state given symbol probability matrix. A copy of the matrix is made.

Parameters

[in] b the new state given symbol probability matrix

Definition at line 91 of file CRNDiscreteHMM.cpp.

void DiscreteHMM::SetStateTransitionProbability ( const SquareMatrixDouble & a )

Sets the state transition probability matrix.

Set the state transition probability matrix. A copy of the matrix is made.

Parameters

[in] a the new state transition probability matrix

Definition at line 80 of file CRNDiscreteHMM.cpp.

String DiscreteHMM::ToString ( ) const

Dumps a summary to a string.

Dump a summary to a string

Returns: a string containing a summary of the model

Definition at line 766 of file CRNDiscreteHMM.cpp.

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

/home/travis/build/Liris-Pleiad/libcrn/src/libcrn/CRNAI/CRNDiscreteHMM.h
/home/travis/build/Liris-Pleiad/libcrn/src/libcrn/CRNAI/CRNDiscreteHMM.cpp

Public Member Functions

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation