Estimation of stopping times for stopped self-similar random processes (Q2046303)

Let \(Y= \{Y_t,t\geq 0\}\) be a self-similar process with scaling parameter \(H.\) Let \(T\) be a non-negative stopping time with density \(f_T\) independent of the process \(Y.\) Let \(X_1,\dots,X_n\) be an i.i.d. sample of \(Y_T.\) The author constructs a nonparametric estimator of the density \(f_T\) using Mellin transform techniques based on the fact that the distribution of \(T^HY_1\) is the same as that of \(Y_T.\) The minimax convergence rate of this estimator is calculated for some examples including the case of the Bessel process. Asymptotic normality of the estimator is proved for the case of the Bessel process. Results of \textit{D. Belomestny} and \textit{J. Schoenmakers} [Stochastic Processes Appl. 126, No. 7, 2092--2122 (2016; Zbl 1337.60089)] were extended to multi-dimensional Brownian motion.