B0416
Title: Dependent wild bootstrap for change-point detection in functional time series and random fields
Authors: Martin Wendler - Otto-von-Guericke University Magdeburg (Germany) [presenting]
Abstract: The aim is to construct a test for the hypothesis of stationarity against the alternative of a location shift in a sequence or fields of dependent, Hilbert-space-valued random variables. Robust tests are also considered, generalizing the Wilcoxon-Mann-Whitney 2-sample U-statistics to functional data. Since this class of test statistics does not rely on dimension reduction, the limit distribution provides an infinite-dimensional covariance operator as a parameter, which is difficult to estimate. Because of this, it is discussed how the dependent wild bootstrap can be adapted to random fields and to U-statistics with values in a Hilbert-space.