Viterbi training vs Baum-Welch algorithm

Question

I'm trying to find the most probable path (i.e. sequence of states) on an HMM using the Viterbi algorithm. However, I don't know the transition and emission matrices, which I need to estimate from the observations (data).

To estimate these matrices, which algorithm should I use: the Baum-Welch algorithm or the Viterbi Training algorithm? Why?

In case I should use the Viterbi training algorithm, can anyone provide me a good pseudocode (it's not easy to find)?

Answer 1

The transition matrix is obtained by using Markov chains. The emission matrix depends upon the hidden states that are unobservable. Each of the observed states may have more than one unobservable For example a person may be healthy or feverish depends upon observable states like cold, dizzy and headache. The matrix giving the transition probability of unobservable state to observable state is emission matrix.