Asymptotics for Hidden Markov Models

Abstract:
I will take the most simple HMM with a finite state space and with symbols
emitted in a finite space and study the asymptotics of the maximum
likelihood estimate. Here the asymptotic situation refers to a growing
number of observations along the chain. For the study we need a general
central limit theorem for dependent variables, results on mixing of the
chain, and a fair amount of numerical analysis.