Some new tests for normality based on U-processes

被引：32

作者：

Arcones, MA ^{[1
]}

Wang, YS ^{[1
]}

机构：

[1] SUNY Binghamton, Dept Math Sci, Binghamton, NY 13902 USA

来源：

STATISTICS & PROBABILITY LETTERS | 2006年 / 76卷 / 01期

关键词：

test of normality; U-processes;

D O I：

10.1016/j.spl.2005.07.003

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 present two new tests for normality based on U-processes. These tests improve on the Lilliefors tests. We obtain the consistency and asymptotic null distribution of these tests. We present simulations of the power of these tests and of several classical tests for some fixed alternatives. These simulations show that the presented tests are competitive. (C) 2005 Elsevier B.V. All rights reserved.

引用

页码：69 / 82

页数：14

共 18 条

[1] LIMIT-THEOREMS FOR U-PROCESSES [J].

ARCONES, MA ;

GINE, E .

ANNALS OF PROBABILITY, 1993, 21 (03) :1494-1542

[2]

Baringhaus L., 1988, METRIKA, V35, P339, DOI DOI 10.1007/BF02613322

[3]

CSORGO M, 1973, J ROY STAT SOC B MET, V35, P507

[4] TESTING FOR NORMALITY IN ARBITRARY DIMENSION [J].

CSORGO, S .

ANNALS OF STATISTICS, 1986, 14 (02) :708-723

[5]

de la Pena V.H., 1999, Decoupling, From dependence to independence, Randomly stopped processes. U-statistics and processes. Martingales and beyond, Probability and its Applications (New York)

[6]

del Barrio E, 2000, TEST-SPAIN, V9, P1

[7]

Dudley RM, 1999, Cambridge Studies in Advanced Mathematics, V63

[8] A TEST FOR NORMALITY BASED ON THE EMPIRICAL CHARACTERISTIC FUNCTION [J].

EPPS, TW ;

PULLEY, LB .

BIOMETRIKA, 1983, 70 (03) :723-726

[9] A CLASS OF INVARIANT CONSISTENT TESTS FOR MULTIVARIATE NORMALITY [J].

HENZE, N ;

ZIRKLER, B .

COMMUNICATIONS IN STATISTICS-THEORY AND METHODS, 1990, 19 (10) :3595-3617

[10]

Lee AJ, 1990, U STAT THEORY PRACTI

← 1 2 →