bwd_seq_gen: Calculate the order in which nodes in the dag should be...

In dagHMM: Directed Acyclic Graph HMM with TAN Structured Emissions

Calculate the order in which nodes in the dag should be traversed during the backward pass(leaves to roots)

Description

dag is a complex graphical model where we can have multiple parents and multiple children for a node. Hence the order in which the dag should be tranversed becomes significant. Backward algorithm is a dynamic programming problem where to calculate the values at a node, we need the values of the children nodes beforehand, which need to be traversed before this node. This algorithm outputs a possible(not unique) order of the traversal of nodes ensuring that the childrens are traversed first before the parents

Usage

bwd_seq_gen(hmm, nlevel = 100)

Arguments

hmm	hmm Object of class List given as output by initHMM
nlevel	No. of levels in the dag, if known. Default is 100

Value

Vector of length "D", where "D" is the number of nodes in the dag

See Also

backward

Examples

library(bnlearn)

tmat = matrix(c(0,0,1,0,0,0,0,1,0,0,0,0,0,1,1,0,0,0,0,0,0,0,0,0,0),
               5,5, byrow= TRUE ) #for "X" (5 nodes) shaped dag
states = c("P","N") #"P" represent cases(or positive) and "N" represent controls(or negative)
bnet = model2network("[A][C|A:B][D|A:C][B|A]") #A is the target variable while
                                               #B, C and D are covariates
obsvA=data.frame(list(B=c("L","H","H","L","L"),C=c("H","H","L","L","H"),D=c("L","L","L","H","H")))
hmmA = initHMM(States=states, dagmat= tmat, net=bnet, observation=obsvA)
bt_sq = bwd_seq_gen(hmmA)

dagHMM documentation built on Jan. 11, 2023, 1:13 a.m.