C# Class Accord.Statistics.Models.Markov.Hybrid.HybridMarkovModel

Hybrid Markov model for arbitrary state-observation functions.

This class can be used to implement HMM hybrids such as ANN-HMM or SVM-HMMs through the specification of a custom GeneralMarkovFunction.

Show file Open project: accord-net/framework

Public Methods

Method	Description
Decode ( double observations, double &logLikelihood ) : int[]	Calculates the most likely sequence of hidden states that produced the given observation sequence. Decoding problem. Given the HMM M = (A, B, pi) and the observation sequence O = {o1,o2, ..., oK}, calculate the most likely sequence of hidden states Si that produced this observation sequence O. This can be computed efficiently using the Viterbi algorithm.
Evaluate ( double observations ) : double	Calculates the probability that this model has generated the given sequence. Evaluation problem. Given the HMM M = (A, B, pi) and the observation sequence O = {o1, o2, ..., oK}, calculate the probability that model M has generated sequence O. This can be computed efficiently using the Forward algorithm.
HybridMarkovModel ( GeneralMarkovFunction function, int states, int dimension ) : System	Initializes a new instance of the HybridMarkovModel class.

Method Details

Decode() public method

Calculates the most likely sequence of hidden states that produced the given observation sequence.

Decoding problem. Given the HMM M = (A, B, pi) and the observation sequence O = {o1,o2, ..., oK}, calculate the most likely sequence of hidden states Si that produced this observation sequence O. This can be computed efficiently using the Viterbi algorithm.

public Decode ( double observations, double &logLikelihood ) : int[]
observations	double	/// A sequence of observations.
logLikelihood	double	/// The state optimized probability.
return	int[]

Evaluate() public method

Calculates the probability that this model has generated the given sequence.

Evaluation problem. Given the HMM M = (A, B, pi) and the observation sequence O = {o1, o2, ..., oK}, calculate the probability that model M has generated sequence O. This can be computed efficiently using the Forward algorithm.

public Evaluate ( double observations ) : double
observations	double	/// A sequence of observations.
return	double

HybridMarkovModel() public method

Initializes a new instance of the HybridMarkovModel class.

public HybridMarkovModel ( GeneralMarkovFunction function, int states, int dimension ) : System
function	GeneralMarkovFunction	A function specifying a probability for a transition-emission pair.
states	int	The number of states in the model.
dimension	int	The number of dimensions in the model.
return	System