pgm.inference package

Submodules

pgm.inference.Gibbs module

class pgm.inference.Gibbs.GibbsSampler(local_distribution, burninT, Xnodes, Ynodes)

Bases: object

Proposal(ii, jj)

count(x)

countUtils(x, c, i, j)

getConditional(x, i, j)

localSampler(x_next, ii, jj)

sampler()

pgm.inference.MetropolisHastings module

pgm.inference.check module

pgm.inference.loopyBP module

pgm.inference.messageParsing module

class pgm.inference.messageParsing.MessageParsing(rootNode, type='BN')

Bases: object

message parsing algorithm

KF Book: chapter 7

parse_message(node1, node2)

Caclulates the message from node1 -> node2 m_{node1->node2} =

sum_{node1} phi(node1)phi(node1, node2)

prod_{pa in parents_{node1}} message_{pa->node1}

node1: [‘int’ or ‘str’] or <Node> object node2: [‘int’ or ‘str’] or <Node> object

Module contents