Sharp error terms for return time statistics under mixing conditions (Q1014064)

The authors' aim is to describe the return time statistics of a string of symbols in a mixing stationary stochastic process with finite alphabet. Let \(\tau _{A}\) denote the time elapsed until the process spells a finite string \(A\) and \(S_{A}\) the number of consecutive repetitions of \(A\). Assuming that the length of the string grows unboundedly, it is proved that (1) the distribution of \(\tau _{A}\), when the process starts with the string \(A\), is well approximated by a certain mixture of the point measure at the origin and an exponential law, and (2) \(S_{A}\) is approximately geometrically distributed. Sharp error terms for each of these approximations are provided. The errors obtained are pointwise and also allow to get approximations for all moments of \(\tau _{A}\) and \(S_{A}\). To obtain (1) it is assumed that the process is \(\phi \)-mixing, while to obtain (2) just the convergence of certain conditional probabilities is required.