gov.sandia.cognition.statistics.bayesian

Interface MetropolisHastingsAlgorithm.Updater<ObservationType,ParameterType>

Type Parameters:

ObservationType - Type of observations handled by the MCMC algorithm.

ParameterType - Type of parameters to infer.

All Superinterfaces:

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

Enclosing class:

MetropolisHastingsAlgorithm<ObservationType,ParameterType>

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

Creates proposals for the MCMC steps.

Method Summary

Modifier and Type	Method and Description
double	computeLogLikelihood(ParameterType parameter, java.lang.Iterable<? extends ObservationType> data) Computes the log likelihood of the data given the parameter
ParameterType	createInitialParameter() Creates the initial parameterization
WeightedValue<ParameterType>	makeProposal(ParameterType location) Makes a proposal update given the current parameter set

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

clone

Method Detail

createInitialParameter

ParameterType createInitialParameter()

Creates the initial parameterization

Returns:

Initial parameters

computeLogLikelihood

double computeLogLikelihood(ParameterType parameter,
                            java.lang.Iterable<? extends ObservationType> data)

Computes the log likelihood of the data given the parameter

Parameters:

parameter - Parameter to consider

data - Data to consider

Returns:

log likelihood of the data given the parameter

makeProposal

WeightedValue<ParameterType> makeProposal(ParameterType location)

Makes a proposal update given the current parameter set

Parameters:

location - Location from which to make a proposal

Returns:

Location of the proposed sample, weighted by the "q" ratio.