Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
1150522 | Journal of Statistical Planning and Inference | 2008 | 25 Pages |
Abstract
We study an autoregressive time series model with a possible change in the regression parameters. Approximations to the critical values for change-point tests are obtained through various bootstrapping methods. Theoretical results show that the bootstrapping procedures have the same limiting behavior as their asymptotic counterparts discussed in Hušková et al. [2007. On the detection of changes in autoregressive time series, I. Asymptotics. J. Statist. Plann. Inference 137, 1243–1259]. In fact, a small simulation study illustrates that the bootstrap tests behave better than the original asymptotic tests if performance is measured by the αα- and ββ-errors, respectively.
Related Topics
Physical Sciences and Engineering
Mathematics
Applied Mathematics
Authors
Marie Hušková, Claudia Kirch, Zuzana Prášková, Josef Steinebach,