gov.sandia.cognition.statistics.bayesian

Interface ParticleFilter.Updater<ObservationType,ParameterType>

Type Parameters:

ObservationType - Type of observations.

ParameterType - Type of parameter to update.

All Superinterfaces:

java.lang.Cloneable, CloneableSerializable, java.io.Serializable

Enclosing interface:

ParticleFilter<ObservationType,ParameterType>

public static interface ParticleFilter.Updater<ObservationType,ParameterType>
extends CloneableSerializable

Updates the particles.

Method Summary

Modifier and Type	Method and Description
double	computeLogLikelihood(ParameterType particle, ObservationType observation) Computes the log likelihood of the parameter and the observation.
DataDistribution<ParameterType>	createInitialParticles(int numParticles) Creates the initial particles.
ParameterType	update(ParameterType previousParameter) Makes a proposal update given the current parameter set

Methods inherited from interface gov.sandia.cognition.util.CloneableSerializable

clone

Method Detail

update

ParameterType update(ParameterType previousParameter)

Makes a proposal update given the current parameter set

Parameters:

previousParameter - Parameters from which to update

Returns:

Proposed parameters

createInitialParticles

DataDistribution<ParameterType> createInitialParticles(int numParticles)

Creates the initial particles.

Parameters:

numParticles - Number of particles to create.

Returns:

Initial particle distribution.

computeLogLikelihood

double computeLogLikelihood(ParameterType particle,
                            ObservationType observation)

Computes the log likelihood of the parameter and the observation.

Parameters:

particle - Parameter to evaluate.

observation - Observation to compute the likelihood of.

Returns:

Log likelihood of the parameter and the observation.