Adjusted Viterbi training for hidden Markov models
Alexey Koloydenko and Jüri Lember
Eurandom, Eindhoven, The Netherlands.
We consider estimation of the emission parameters in hidden Markov models. Commonly, one uses the EM algorithm for this purpose. However, our primary motivation is the Philips speech recognition system wherein the EM algorithm is replaced by the Viterbi training algorithm. Viterbi training is faster and computationally less involved than EM, but it is also biased and need not even be consistent. For this reason we propose an alternative to the Viterbi training -- adjusted Viterbi training -- that has the same order of computational complexity as Viterbi training but gives more accurate estimators. Elsewhere, we studied the adjusted Viterbi training for a special case of mixtures with relevant simulations ascertaining the theory. This paper shows how the adjusted Viterbi training is also possible for more general hidden Markov models.