recently i'm doing training of hmm, used hmm toolbox. have problems , couldn't resolve them.
i train hmm shown below. there's no problems here.
[ll, prior1, transmatrix1, observematrix1] = dhmm_em(data, prior0, transmatrix0, observematrix0);i use viterbi algorithm find most-probable path through hmm state trellis.
function path = viterbi_path(prior, transmat, obslik);now there's problem. don't know "obslik" means. observematrix1?
i want probability of sequence, don't know whether should use "fwdback" function or not. if should, "obslik" means then?
function [alpha, beta, gamma, loglik, xi_summed, gamma2] = fwdback(init_state_distrib, transmat, obslik, varargin);thanks!!!
i didn't understand comments. understand it.
the "obslik" here isn't equal observematrix1. before using viterbi_path function, should compute obslik:
obslik = multinomial_prob(data(m,:), observematrix1); data matrix observematrix0, observe-matrix before training. am right?
