baumWelchRecursion: Implementation of the Baum Welch Algorithm as a special case...
In treeHMM: Tree Structured Hidden Markov Model

Description Usage Arguments Value See Also Examples

View source: R/baumWelch.R

baumWelch recursively calls this function to give a final estimate of parameters for tree HMM Uses Parallel Processing to speed up calculations for large data. Should not be used directly.

1 2	baumWelchRecursion(hmm, observation, kn_states = NULL, kn_verify = NULL)

`hmm`	hmm Object of class List given as output by `initHMM`
`observation`	A list consisting "k" vectors for "k" features, each vector being a character series of discrete emmision values at different nodes serially sorted by node number
`kn_states`	(Optional) A (L * 2) dataframe where L is the number of training nodes where state values are known. First column should be the node number and the second column being the corresponding known state values of the nodes
`kn_verify`	(Optional) A (L * 2) dataframe where L is the number of validation nodes where state values are known. First column should be the node number and the second column being the corresponding known state values of the nodes

List containing estimated Transition and Emission probability matrices

baumWelch

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 tree
hmmA = initHMM(c("P","N"),list(c("L","R")), tmat) #one feature with two discrete levels "L" and "R"
obsv = list(c("L","L","R","R","L")) #emissions for the one feature for the 5 nodes in order 1:5
kn_st = data.frame(node=c(2),state=c("P"),stringsAsFactors = FALSE)
                   #state at node 2 is known to be "P"
kn_vr = data.frame(node=c(3,4,5),state=c("P","N","P"),stringsAsFactors = FALSE) 
                   #state at node 3,4,5 are "P","N","P" respectively
newparam= baumWelchRecursion(hmmA,obsv,kn_st, kn_vr)