Adaptive encoding and prediction of hidden Markov processes
László Gerencsér, Gábor Molnár-Sáska · 2003
The purpose of this paper is to provide explicit results on the almost sure asymptotic performance of adaptive encoding and prediction procedures for finite-state Hidden Markov Models. In addition, Rissanen's tail condition [14] will be verified, from which a lower bound for the mean-performance of universal encoding procedures will be derived. The results of this paper are based on [10].