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

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

Classes
class	BackSamplingDistribution
Public Member Functions
	BackwardSamplingWithChildren (BeliefNetworkEx bn) throws Exception
SampledDistribution	infer () throws Exception
void	prepareInference (int[] evidenceDomainIndices) throws Exception
Protected Member Functions
BackSamplingDistribution	getBackSamplingDistribution (BeliefNode node, WeightedSample s)
Protected Attributes
Cache2D< BeliefNode, Long, BackSamplingDistribution >	distCache
Stopwatch	distSW
Cache2D< CPF, Integer, Double >	probCache
Stopwatch	probSW

Detailed Description

a backward sampling algorithm that, to sample the parents of an instantiated node N, considers not only the conditional probability of N given its parents but also the the children of N's parents and their parents (using existing instantiations and, where nodes are yet uninstantiated, the prior probability of the nodes)

Author:: jain

Definition at line 23 of file BackwardSamplingWithChildren.java.

Constructor & Destructor Documentation

edu::tum::cs::bayesnets::inference::BackwardSamplingWithChildren::BackwardSamplingWithChildren ( BeliefNetworkEx bn ) throws Exception [inline]

Definition at line 241 of file BackwardSamplingWithChildren.java.

Member Function Documentation

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