Leroux's method for general hidden Markov models (Q2490057): Difference between revisions

The authors consider a general hidden Markov model where the state variable evolves in an open interval of the real line. Using \textit{B. G. Leroux}'s method [Stochastic Processes Appl. 40, No. 1, 127--143 (1992; Zbl 0738.62081)], they prove under rather minimal assumptions the convergence of the normalized log-likelihood to the value of the parameter. The example of Kalman filter model is studied in full details.