Almost sure convergence of the kT-occupation time density estimator

In this Note, we introduce an extension of the k-nearest neighbor estimator in continuous time, the $k_T$-occupation time estimator, and we give sufficient conditions for its existence. Then, we show the almost sure convergence for α-mixing and bounded processes in two cases, the superoptimal case (when parametric rates are reached) and the optimal case (when i.i.d. rates of density estimation are reached). To cite this article: B. Labrador, C. R. Acad. Sci. Paris, Ser. I 343 (2006).