Bootstrap confidence regions in multinomial sampling

被引：9
|
作者：
Morales, D
Pardo, L
Santamaría, L
机构：
[1] Miguel Hernandez Univ Elche, Ctr Operat Res, Elche, Spain
[2] Univ Complutense, Dept Stat & OR, E-28040 Madrid, Spain
来源：
APPLIED MATHEMATICS AND COMPUTATION | 2004年 / 155卷 / 02期
关键词：
power divergences; bootstrap; confidence regions; average coverage probability; Monte Carlo simulation;
D O I：
10.1016/S0096-3003(03)00777-X
中图分类号：
O29 [应用数学];
学科分类号：
070104 ;
摘要：
Power divergences can be used to give a measure of distance between two probability vectors. In multinomial sampling arguments can be substituted by empirical and theoretical proportions to obtain confidence regions of parameters. In this paper the bootstrap versions of these confidence regions are constructed. Monte Carlo simulation experiments are carried out to calculate average coverage probabilities and to compare the behavior of the introduced procedures. (C) 2003 Elsevier Inc. All rights reserved.
引用
收藏
页码：295 / 315
页数：21
相关论文
共 50 条
12345