Compound Poisson approximation for Markov chains using Stein's method (Q1807200)

Let \(\eta\) be a stationary Harris recurrent Markov chain on a Polish state space \((S,{\mathcal F})\) with stationary distribution \(\mu\). Let \(\Psi_n:= \sum^n_{i=1} I\{\eta_i\in S_1\}\) be the number of visits to \(S_1\in{\mathcal F}\) by \(\eta\), where \(S_1\) is rare, in the sense that \(\mu(S_1)\) is ``small'', the distribution of \(\Psi_n\) is approximated in total variation by a compound Poisson distribution, in a natural way which takes into account the regenerative properties of Harris recurrent Markov chains. When the chain has an atom \(S_0\) such that \(\mu(S_0)> 0\), the bound depends only on much studied quantities like hitting probabilities and expected hitting times, which satisfy Poisson's equation. The results are illustrated by numerical evaluations of the error bound for some Markov chains on finite state spaces.