On the asymptotic properties of a nonparametric L1-test statistic of homogeneity

被引:31
|
作者
Biau, G
Györfi, L
机构
[1] Univ Montpellier 2, CNRS, UMR 5149,Equipe Probabilites & Stat, Inst Math & Modelisat Montpellier, F-34095 Montpellier, France
[2] Budapest Univ Technol & Econ, Dept Comp Sci & Informat Theory, H-1521 Budapest, Hungary
关键词
central limit theorem; consistent testing; homogeneity testing; large deviations; partitions; poissonization;
D O I
10.1109/TIT.2005.856979
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
We present two simple and explicit procedures for testing homogeneity of two independent multivariate samples of size n. The non-parametric tests are based on the statistic T-n, which is the L-1 distance between the two empirical distributions restricted to a finite partition. Both tests reject the null hypothesis of homogeneity if T-n becomes large, i.e., if T-n exceeds a threshold. We first discuss Chernoff-type large deviation properties of T-n. This results in a distribution-free strong consistent test of homogeneity. Then the asymptotic null distribution of the test statistic is obtained, leading to an asymptotically alpha-level test procedure.
引用
收藏
页码:3965 / 3973
页数:9
相关论文
共 50 条