Some recent extreme value asymptotics for multivariate renewal processes are used to derive an asymptotic changepoint test. This test is proven to be consistent in the multivariate framework where we assume that at most one change (AMOC) occurrs in any of the component renewal processes. Since the actual covariance structure is often unknown, we also suggest an appropriate estimate.
KeywordsCentral Limit Theorem Renewal Process Invariance Principle Local Lipschitz Condition Math Proc
Unable to display preview. Download preview PDF.
- Csörgő M, Horváth L (1993) Weighted approximations in probability and statistics. Wiley, ChichesterGoogle Scholar
- Csörgő M, Révész P (1981) Strong approximations in probability and statistics. Academic Press, New YorkGoogle Scholar
- Eastwood VR, Steinebach J, Warren J (1996) A simulation study on a test of change in a multivariate renewal process. Tech Rep Ser Dept Stat, Univ Auckland (in progress)Google Scholar
- Eastwood VR, Warren J (1993) Simulations of Kendall-Kendall pontograms and related point processes. Tech Rep Ser Lab Res Stat Probab, Carleton Univ Ottawa 229:1–131Google Scholar
- Golub GH, van Loan CF (1989) Matrix computations. John Hopkins Univ Press (2nd ed)Google Scholar
- Horváth L (1993) The maximum likelihood method for testing changes in the parameters of normal observations. Ann Statist 21:20–68Google Scholar
- Steinebach J (1994) Change point and jump estimates in an AMOC renewal model. In: Mandl P, Hušková M (Eds) Asymptotic Statistics, Proceedings, Fifth Prague Symposium, Sep 4–9, 1993, p 447–457. Contributions to Statistics. Physica-Verlag, HeidelbergGoogle Scholar