be.ac.ulg.montefiore.run.jahmm
Class OpdfMultiGaussian

java.lang.Object
  be.ac.ulg.montefiore.run.jahmm.OpdfMultiGaussian

All Implemented Interfaces:

Opdf<ObservationVector>, java.io.Serializable, java.lang.Cloneable

public class OpdfMultiGaussian
extends java.lang.Object
implements Opdf<ObservationVector>

This class represents a multivariate gaussian distribution function.

See Also:

Serialized Form

Constructor Summary

OpdfMultiGaussian(double[] mean, double[][] covariance)
Builds a new gaussian probability distribution with a given mean and covariance matrix.

OpdfMultiGaussian(int dimension)
Builds a new gaussian probability distribution with zero mean and identity covariance matrix.

Method Summary

OpdfMultiGaussian	clone()
double[][]	covariance() Returns (a copy of) this distribution's covariance matrix.
int	dimension() Returns the dimension of the vectors handled by this distribution.
void	fit(java.util.Collection<? extends ObservationVector> co) Fits this observation probability (distribution) function to a (non empty) set of observations.
void	fit(java.util.Collection<? extends ObservationVector> co, double[] weights) Fits this observation probability (distribution) function to a weighted (non empty) set of observations.
void	fit(ObservationVector... oa) Fits this observation probability (distribution) function to a (non empty) set of observations.
void	fit(ObservationVector[] o, double[] weights) Fits this observation probability (distribution) function to a weighted (non empty) set of observations.
ObservationVector	generate() Generates a (pseudo) random observation according to this distribution.
double[]	mean() Returns (a copy of) this distribution's mean vector.
double	probability(ObservationVector o) Returns the probability (density) of an observation given a distribution.
java.lang.String	toString()
java.lang.String	toString(java.text.NumberFormat numberFormat) Returns a String describing this distribution.

Methods inherited from class java.lang.Object

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

Constructor Detail

OpdfMultiGaussian

public OpdfMultiGaussian(int dimension)

Builds a new gaussian probability distribution with zero mean and identity covariance matrix.

Parameters:

dimension - The dimension of the vectors.

OpdfMultiGaussian

public OpdfMultiGaussian(double[] mean,
                         double[][] covariance)

Builds a new gaussian probability distribution with a given mean and covariance matrix.

Parameters:

mean - The distribution's mean.

covariance - The distribution's covariance matrix.

Method Detail

mean

public double[] mean()

Returns (a copy of) this distribution's mean vector.

Returns:

The mean vector.

covariance

public double[][] covariance()

Returns (a copy of) this distribution's covariance matrix.

Returns:

The covariance matrix.

dimension

public int dimension()

Returns the dimension of the vectors handled by this distribution.

Returns:

The dimension of the vectors handled by this distribution.

probability

public double probability(ObservationVector o)

Description copied from interface: Opdf

Returns the probability (density) of an observation given a distribution.

Specified by:

probability in interface Opdf<ObservationVector>

Parameters:

o - An observation.

Returns:

The probability (density, if o takes continuous values) of o for this function.

generate

public ObservationVector generate()

Description copied from interface: Opdf

Generates a (pseudo) random observation according to this distribution.

Specified by:

generate in interface Opdf<ObservationVector>

Returns:

An observation.

fit

public void fit(ObservationVector... oa)

Description copied from interface: Opdf

Fits this observation probability (distribution) function to a (non empty) set of observations. The meaning to give to fits should be has the maximum likelihood if possible.

Specified by:

fit in interface Opdf<ObservationVector>

Parameters:

oa - An array of observations compatible with this function.

fit

public void fit(java.util.Collection<? extends ObservationVector> co)

Description copied from interface: Opdf

Fits this observation probability (distribution) function to a (non empty) set of observations. The meaning to give to fits should be has the maximum likelihood if possible.

Specified by:

fit in interface Opdf<ObservationVector>

Parameters:

co - A set of observations compatible with this function.

fit

public void fit(ObservationVector[] o,
                double[] weights)

Description copied from interface: Opdf

Fits this observation probability (distribution) function to a weighted (non empty) set of observations. Equations (53) and (54) of Rabiner's A Tutorial on Hidden Markov Models and Selected Applications in Speech Recognition explain how the weights can be used.

Specified by:

fit in interface Opdf<ObservationVector>

Parameters:

o - An array of observations compatible with this factory.

weights - The weight associated to each observation (such that weight.length == o.length and the sum of all the elements equals 1).

fit

public void fit(java.util.Collection<? extends ObservationVector> co,
                double[] weights)

Description copied from interface: Opdf

Fits this observation probability (distribution) function to a weighted (non empty) set of observations. Equations (53) and (54) of Rabiner's A Tutorial on Hidden Markov Models and Selected Applications in Speech Recognition explain how the weights can be used.

Specified by:

fit in interface Opdf<ObservationVector>

Parameters:

co - A set of observations compatible with this factory.

weights - The weight associated to each observation (such that weight.length == o.length and the sum of all the elements equals 1).

clone

public OpdfMultiGaussian clone()

Specified by:

clone in interface Opdf<ObservationVector>

Overrides:

clone in class java.lang.Object

toString

public java.lang.String toString()

Overrides:

toString in class java.lang.Object

toString

public java.lang.String toString(java.text.NumberFormat numberFormat)

Description copied from interface: Opdf

Returns a String describing this distribution.

Specified by:

toString in interface Opdf<ObservationVector>

Parameters:

numberFormat - A formatter used to convert the numbers (e.g. probabilities) to strings.

Returns:

A String describing this distribution.