Title: Statistical inference for clusters of extremes: Disjoint vs. sliding blocks estimators
Authors: Rafal Kulik - University of Ottawa (Canada) [presenting]
Abstract: Limit theorems for empirical cluster functionals are presented. We consider both disjoint and sliding block estimators. Conditions for consistency and asymptotic normality are provided in terms of tail and spectral tail processes and can be verified for a large class of multivariate time series, including geometrically ergodic Markov chains. Applications include asymptotic normality for the classical extremal index and recently introduced cluster indices. Results for multiplier bootstrap processes are also provided.