B0938
Title: Detecting early or late change-points in time series using U-statistics
Authors: Kata Vuk - University of Regensburg (Germany) [presenting]
Herold Dehling - Ruhr-University Bochum (Germany)
Martin Wendler - Otto-von-Guericke University Magdeburg (Germany)
Abstract: The focus is on non-parametric weighted change-point tests that are based on two-sample U-statistics. By a suitable choice of weights, one obtains tests that are able to detect changes in time series that occur very early or late during the observation period. The limit distribution of those test statistics is investigated under the hypothesis that no change occurs, but also under the alternative that there is a change in mean. To illustrate the results, some simulations and applications to real-life data will be presented.
