dev.nm.stat.hmm.mixture.distribution

Class LogNormalMixtureDistribution

java.lang.Object

dev.nm.stat.hmm.mixture.distribution.LogNormalMixtureDistribution

All Implemented Interfaces:

MixtureDistribution

public class LogNormalMixtureDistribution
extends Object
implements MixtureDistribution

The HMM states use the Log-Normal distribution to model the observations.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	LogNormalMixtureDistribution.Lambda the log-normal distribution parameters

Constructor Summary

Constructors

Constructor and Description
LogNormalMixtureDistribution(LogNormalMixtureDistribution.Lambda[] lambda) Constructs a log-normal distribution for each state in the HMM model.
LogNormalMixtureDistribution(LogNormalMixtureDistribution.Lambda[] lambda, boolean isMuEstimated, boolean isSigmaEstimated) Constructs a log-normal distribution for each state in the HMM model.

Method Summary

Modifier and Type	Method and Description
LogNormalMixtureDistribution.Lambda[]	getMStepParams(double[] observations, Vector[] u) Maximize, for each state, the log-likelihood of the distribution with respect to the observations and current estimators.
LogNormalMixtureDistribution.Lambda[]	getParams() Get the parameters, for each state, of the distribution.
ProbabilityDistribution[]	newDistributions() Get the distributions (possibly differently parameterized) for all states.
MixtureDistribution	newMixtureDistribution(Object[] param) Construct a new distribution from a set of parameters, one set per state.
RandomNumberGenerator[]	newRandomNumberGenerators() Get the random number generators corresponding to the distributions (possibly differently parameterized) for all states.

Methods inherited from class java.lang.Object

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

Constructor Detail

LogNormalMixtureDistribution

public LogNormalMixtureDistribution(LogNormalMixtureDistribution.Lambda[] lambda,
                                    boolean isMuEstimated,
                                    boolean isSigmaEstimated)

Constructs a log-normal distribution for each state in the HMM model.

Parameters:

lambda - the distribution parameters

isMuEstimated - indicate whether parameter mu is to be estimated

isSigmaEstimated - indicate whether parameter sigma is to be estimated

LogNormalMixtureDistribution

public LogNormalMixtureDistribution(LogNormalMixtureDistribution.Lambda[] lambda)

Constructs a log-normal distribution for each state in the HMM model.

Parameters:

lambda - the distribution parameters

Method Detail

getParams

public LogNormalMixtureDistribution.Lambda[] getParams()

Description copied from interface: MixtureDistribution

Get the parameters, for each state, of the distribution.

Specified by:

getParams in interface MixtureDistribution

Returns:

the parameters, for each state, of the distribution

newRandomNumberGenerators

public RandomNumberGenerator[] newRandomNumberGenerators()

Description copied from interface: MixtureDistribution

Get the random number generators corresponding to the distributions (possibly differently parameterized) for all states.

Specified by:

newRandomNumberGenerators in interface MixtureDistribution

Returns:

the random number generators corresponding to the distributions

newDistributions

public ProbabilityDistribution[] newDistributions()

Description copied from interface: MixtureDistribution

Get the distributions (possibly differently parameterized) for all states.

Specified by:

newDistributions in interface MixtureDistribution

Returns:

the distributions

getMStepParams

public LogNormalMixtureDistribution.Lambda[] getMStepParams(double[] observations,
                                                            Vector[] u)

Description copied from interface: MixtureDistribution

Maximize, for each state, the log-likelihood of the distribution with respect to the observations and current estimators.

Specified by:

getMStepParams in interface MixtureDistribution

Parameters:

observations - the observations

u - this is in eq. (4.13) of Zucchini and MacDonald (2009), p. 65

Returns:

the MLE estimators

newMixtureDistribution

public MixtureDistribution newMixtureDistribution(Object[] param)

Description copied from interface: MixtureDistribution

Construct a new distribution from a set of parameters, one set per state.

Specified by:

newMixtureDistribution in interface MixtureDistribution

Parameters:

param - the parameters of the distribution, one set per state

Returns:

a new HMM distribution