edu::tum::cs::bayesnets::inference::BackwardSampling Class Reference

Inheritance diagram for edu::tum::cs::bayesnets::inference::BackwardSampling:

[legend]

List of all members.

Classes
class	BackSamplingDistribution
class	TierComparator
Public Member Functions
SampledDistribution	_infer () throws Exception
	BackwardSampling (BeliefNetworkEx bn) throws Exception
void	getSample (WeightedSample s) throws Exception
void	initSample (WeightedSample s) throws Exception
Protected Member Functions
BackSamplingDistribution	getBackSamplingDistribution (BeliefNode node, WeightedSample s)
void	getOrdering (int[] evidenceDomainIndices) throws Exception
void	onAddedSample (WeightedSample s) throws Exception
void	prepareInference (int[] evidenceDomainIndices) throws Exception
boolean	sampleBackward (BeliefNode node, WeightedSample s)
boolean	sampleForward (BeliefNode node, WeightedSample s)
Protected Attributes
Vector< BeliefNode >	backwardSampledNodes
int	currentStep
Vector< BeliefNode >	forwardSampledNodes
HashSet< BeliefNode >	outsideSamplingOrder

---

Detailed Description

an implementation of the backward simulation algorithm as described by Robert Fung and Brendan Del Favero in "Backward Simulation in Bayesian Networks" (UAI 1994)

Author:

jain

Definition at line 24 of file BackwardSampling.java.

---

Constructor & Destructor Documentation

edu::tum::cs::bayesnets::inference::BackwardSampling::BackwardSampling

(

BeliefNetworkEx

bn

)

throws Exception [inline]

Definition at line 92 of file BackwardSampling.java.

---

Member Function Documentation

SampledDistribution edu::tum::cs::bayesnets::inference::BackwardSampling::_infer

(

)

throws Exception [inline, virtual]

Implements edu::tum::cs::bayesnets::inference::Sampler.

Reimplemented in edu::tum::cs::srl::bayesnets::inference::LiftedBackwardSampling::Sampler.

Definition at line 210 of file BackwardSampling.java.

BackSamplingDistribution edu::tum::cs::bayesnets::inference::BackwardSampling::getBackSamplingDistribution	(	BeliefNode	node,
		WeightedSample	s
	)			[inline, protected]

Reimplemented in edu::tum::cs::bayesnets::inference::BackwardSamplingWithChildren, edu::tum::cs::bayesnets::inference::BackwardSamplingWithPriors, and edu::tum::cs::srl::bayesnets::inference::LiftedBackwardSampling::Sampler.

Definition at line 193 of file BackwardSampling.java.

void edu::tum::cs::bayesnets::inference::BackwardSampling::getOrdering

(

int[]

evidenceDomainIndices

)

throws Exception [inline, protected]

gets the sampling order by filling the members for backward and forward sampled nodes as well as the set of nodes not in the sampling order

Parameters:

evidenceDomainIndices

Exceptions:

Exception

Reimplemented in edu::tum::cs::bayesnets::inference::SATIS_BSampler.

Definition at line 119 of file BackwardSampling.java.

void edu::tum::cs::bayesnets::inference::BackwardSampling::getSample

(

WeightedSample

s

)

throws Exception [inline]

gets one full sample of all of the nodes

Parameters:

s

Exceptions:

Exception

Definition at line 239 of file BackwardSampling.java.

void edu::tum::cs::bayesnets::inference::BackwardSampling::initSample

(

WeightedSample

s

)

throws Exception [inline]

Reimplemented in edu::tum::cs::bayesnets::inference::SATIS_BSampler.

Definition at line 303 of file BackwardSampling.java.

void edu::tum::cs::bayesnets::inference::BackwardSampling::onAddedSample

(

WeightedSample

s

)

throws Exception [inline, protected]

Reimplemented in edu::tum::cs::srl::bayesnets::inference::SATISExGibbs::SATIS_BSampler_Gibbs.

Definition at line 316 of file BackwardSampling.java.

void edu::tum::cs::bayesnets::inference::BackwardSampling::prepareInference

(

int[]

evidenceDomainIndices

)

throws Exception [inline, protected]

Reimplemented in edu::tum::cs::bayesnets::inference::BackwardSamplingWithChildren, edu::tum::cs::bayesnets::inference::BackwardSamplingWithPriors, and edu::tum::cs::srl::bayesnets::inference::LiftedBackwardSampling::Sampler.

Definition at line 199 of file BackwardSampling.java.

boolean edu::tum::cs::bayesnets::inference::BackwardSampling::sampleBackward	(	BeliefNode	node,
		WeightedSample	s
	)			[inline, protected]

samples backward from the given node, instantiating its parents

Parameters:

	node
	s	the sample to store the instantiation information in; the weight is also updated with the normalizing constant that is obtained

Returns:

true if sampling succeeded, false otherwise

Definition at line 170 of file BackwardSampling.java.

boolean edu::tum::cs::bayesnets::inference::BackwardSampling::sampleForward	(	BeliefNode	node,
		WeightedSample	s
	)			[inline, protected]

Definition at line 308 of file BackwardSampling.java.

---

Member Data Documentation

Vector<BeliefNode> edu::tum::cs::bayesnets::inference::BackwardSampling::backwardSampledNodes [protected]

Definition at line 26 of file BackwardSampling.java.

int edu::tum::cs::bayesnets::inference::BackwardSampling::currentStep [protected]

Definition at line 29 of file BackwardSampling.java.

Vector<BeliefNode> edu::tum::cs::bayesnets::inference::BackwardSampling::forwardSampledNodes [protected]

Definition at line 27 of file BackwardSampling.java.

HashSet<BeliefNode> edu::tum::cs::bayesnets::inference::BackwardSampling::outsideSamplingOrder [protected]

Definition at line 28 of file BackwardSampling.java.

---

The documentation for this class was generated from the following file:

• BackwardSampling.java