Invariance properties of some classical tests for exponentiality

被引:8
|
作者
Gupta, RD
Richards, DS
机构
[1] UNIV NEW BRUNSWICK,DEPT MATH STAT & COMP SCI,ST JOHNS,NB E2L 4L5,CANADA
[2] UNIV VIRGINIA,DIV STAT,CHARLOTTESVILLE,VA 22903
关键词
characterizations; Dirichlet distribution; empirical distribution function; exponential distribution; Kolmogorov-Smirnov tests; Liouville distribution; reliability theory; robustness of tests; spacings;
D O I
10.1016/S0378-3758(97)00016-5
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
We examine some classical tests for the exponentiality of independent, identically distributed data. We show that a large number of these tests have the same distribution if the data follow certain multivariate Liouville distributions. These results highlight the role that the assumption of independence plays in the behavior of the classical test statistics. We use these results to derive some characterizations of the exponential distributions among the Liouville distributions. (C) 1997 Elsevier Science B.V.
引用
收藏
页码:203 / 213
页数:11
相关论文
共 50 条