io.improbable.keanu.algorithms.mcmc

Class MetropolisHastingsSampler

java.lang.Object

io.improbable.keanu.algorithms.mcmc.MetropolisHastingsSampler

All Implemented Interfaces:

SamplingAlgorithm

public class MetropolisHastingsSampler
extends java.lang.Object
implements SamplingAlgorithm

Constructor Summary

Constructors

Constructor and Description
MetropolisHastingsSampler(java.util.List<? extends Variable> latentVariables, java.util.List<? extends Variable> variablesToSampleFrom, MetropolisHastingsStep mhStep, MHStepVariableSelector variableSelector, double logProbabilityBeforeStep)

Method Summary

Modifier and Type	Method and Description
NetworkSample	sample() Takes a sample with the algorithm and returns the state of the network for that sample.
void	sample(java.util.Map<VariableReference,java.util.List<?>> samplesByVariable, java.util.List<java.lang.Double> logOfMasterPForEachSample) Takes a sample with the algorithm and saves it in the supplied map (creating a new entry in the list if the Variable already exists).
void	step() Move forward the state of the Sampling Algorithm by a single step but do not return anything.

Methods inherited from class java.lang.Object

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

Methods inherited from interface io.improbable.keanu.algorithms.mcmc.SamplingAlgorithm

takeSample

Constructor Detail

MetropolisHastingsSampler

public MetropolisHastingsSampler(java.util.List<? extends Variable> latentVariables,
                                 java.util.List<? extends Variable> variablesToSampleFrom,
                                 MetropolisHastingsStep mhStep,
                                 MHStepVariableSelector variableSelector,
                                 double logProbabilityBeforeStep)

Method Detail

step

public void step()

Description copied from interface: SamplingAlgorithm

Move forward the state of the Sampling Algorithm by a single step but do not return anything.

Specified by:

step in interface SamplingAlgorithm

sample

public void sample(java.util.Map<VariableReference,java.util.List<?>> samplesByVariable,
                   java.util.List<java.lang.Double> logOfMasterPForEachSample)

Description copied from interface: SamplingAlgorithm

Takes a sample with the algorithm and saves it in the supplied map (creating a new entry in the list if the Variable already exists).

Specified by:

sample in interface SamplingAlgorithm

Parameters:

samplesByVariable - map to store sampled variable values

logOfMasterPForEachSample - list of log of master probability for each sample

sample

public NetworkSample sample()

Description copied from interface: SamplingAlgorithm

Takes a sample with the algorithm and returns the state of the network for that sample.

Specified by:

sample in interface SamplingAlgorithm

Returns:

a network state that represents the current state of the algorithm.