Steady state coupled hidden Markov model
Hafssa Kroumbi, Abdelaziz Nasroallah · Communications in Statistics - Simulation and Computation · 2025
We focus on the computation of the steady state probability distribution for coupled hidden Markov chains, where both the state space and emission space are finite and bi-dimensional. Primarily, we use the regenerative property of the Markov chain to estimate the steady state for both the Markov chain and the emission process. Furthermore, we study the asymptotic normality of the estimator proposed for the steady-state distribution of emissions. The effectiveness of the proposed estimators is demonstrated through a sample of numerical Monte Carlo examples.