com.numericalmethod.suanshu.stats.hmm.mixture

Class HmmForwardBackward

java.lang.Object

com.numericalmethod.suanshu.stats.hmm.mixture.HmmForwardBackward

public class HmmForwardBackward
extends java.lang.Object

The implementation of the forward-backward algorithm computes the log of forward and backward probabilities, give a sequence of observations and the hmm model.

See Also:

• L. R. Rabiner, "A tutorial on hidden Markov models and selected applications in speech recognition," Proceedings of the IEEE, Volume: 77, Issue:2, 257 - 286, Feb 1989."
• W. Zucchini and I. L. MacDonald, "Hidden Markov Models for Time Series: An Introduction Using R," Boca Raton, Florida, CRC Press, 2009.

Constructor Summary

Constructors

Constructor and Description
HmmForwardBackward(HiddenMarkovModel model) Construct an instance of HmmForwardBackward to compute the forward and backward probabilities and the log-likelihood.

Method Summary

Modifier and Type	Method and Description
Matrix	logBackward(double[] observations) Get the log-transformed backward probability matrix.
Matrix	logForward(double[] observations) Get the log-transformed forward probability matrix.
double	logLikelihood(double[] observations) Get the log-likelihood.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

HmmForwardBackward

public HmmForwardBackward(HiddenMarkovModel model)

Construct an instance of HmmForwardBackward to compute the forward and backward probabilities and the log-likelihood.

Parameters:

model - a mixture hidden Markov model

Method Detail

logForward

public Matrix logForward(double[] observations)

Get the log-transformed forward probability matrix.

Parameters:

observations - a sequence of observations

Returns:

the log-transformed forward probabilities matrix (number of observations * number of states)

logBackward

public Matrix logBackward(double[] observations)

Get the log-transformed backward probability matrix.

Parameters:

observations - a sequence of observations

Returns:

the log-transformed backward probabilities matrix (number of observations * number of states)

logLikelihood

public double logLikelihood(double[] observations)

Get the log-likelihood.

Parameters:

observations - a sequence of observations

Returns:

the log-likelihood