CFE 2015: Start Registration
View Submission - CFE
A0334
Title: Detecting relevant changes in time series models Authors:  Holger Dette - Ruhr-Universitaet Bochum (Germany)
Dominik Wied - University of Cologne (Germany) [presenting]
Abstract: Most of the literature on change point analysis by means of hypothesis testing considers hypotheses of the form ``Is there a change or not?'' A different perspective is taken, i.e., the null hypothesis of no relevant changes of the form ``Is there a small change or not?'' is considered. A general approach to problems of this type is developed which is based on the cumulative sum principle. For the asymptotic analysis weak convergence of the sequential empirical process must be established under the alternative of non-stationarity, and it is shown that the resulting test statistic is asymptotically normally distributed. Applications of the methodology include the mean and the covariance matrix. The finite sample properties of the new tests are investigated by means of a simulation study and illustrated by analyzing a data example from portfolio management.