Robust Change Detection in the Dependence Structure of Multivariate Time Series
- 1.4k Downloads
A robust change-point test based on the spatial sign covariance matrix is proposed. A major advantage of the test is its computational simplicity, making it particularly appealing for robust, high-dimensional data analysis. We derive the asymptotic distribution of the test statistic for stationary sequences, which we allow to be near-epoch dependent in probability (P NED) with respect to an α-mixing process. Contrary to the usual L2 near-epoch dependence, this short-range dependence condition requires no moment assumptions, and includes arbitrarily heavy-tailed processes. Further, we give a short review of the spatial sign covariance matrix and compare our test to a similar one based on the sample covariance matrix in a simulation study.
KeywordsGARCH Near epoch dependence Oja sign covariance matrix Orthogonal invariance Spatial sign covariance matrix Tyler matrix
This work was supported in part by the Collaborative Research Grant 823 of the German Research Foundation.
- Dehling, H., Vogel, D., Wendler, M., Wied, D.: Testing for changes in the rank correlation of time series. version 1 (2012) [arXiv 1203.4871]Google Scholar
- Fischer, D., Möttönen, J., Nordhausen, K., Vogel, D.: OjaNP: Multivariate Methods Based on the Oja Median and Related Concepts, 2014. http://CRAN.R-project.org/package=OjaNP. R package version 0.9-8
- Tukey, J.W.: A survey of sampling from contaminated distributions. In: Olkin, I., Ghurye, S.G., Hoeffding, W., Madow, W.G., Mann, H.B. (eds.) Contributions to Probability and Statistics. Essays in Honor of Harold Hotelling, pp. 448–485. Stanford University Press, Stanford (1960)Google Scholar
- Visuri, S.: Array and multichannel signal processing using nonparametric statistics. Ph.D. Thesis, Helsinki University of Technology, Helsinki, Finland (2001)Google Scholar