Nonparametric tests for change-point detection a la Gombay and Horvath

JOURNAL OF MULTIVARIATE ANALYSIS

卷:115

Article

Holmes, Mark¹  Kojadinovic, Ivan²  Quessy, Jean-Francois³ 

[1] Univ Auckland, Dept Stat, Auckland 1142, New Zealand

[2] Univ Pau & Pays Adour, UMR CNRS 5142, Lab Math & Applicat, F-64013 Pau, France

[3] Univ Quebec Trois Rivieres, Dept Math & Informat, Trois Rivieres, PQ G9A 5H7, Canada

关键词: Half-spaces;    Lower-left orthants;    Multiplier central limit theorem;    Multivariate independent observations;    Partial-sum process;   

DOI  :  10.1016/j.jmva.2012.10.004

来源: Elsevier

【 摘 要 】

The nonparametric test for change-point detection proposed by Gombay and Horvath is revisited and extended in the broader setting of empirical process theory. The resulting testing procedure for potentially multivariate observations is based on a sequential generalization of the functional multiplier central limit theorem and on modifications of Gombay and Horvath's seminal approach that appears to improve the finite-sample behavior of the tests. A large number of candidate test statistics based on processes indexed by lower-left orthants and half-spaces are considered and their performance is studied through extensive Monte Carlo experiments involving univariate, bivariate and trivariate data sets. Finally, practical recommendations are provided and the tests are illustrated on trivariate hydrological data. (C) 2012 Elsevier Inc. All rights reserved.

【 授权许可】

Free