FUNCTIONAL LIMIT-THEOREMS FOR INVERSE BOOTSTRAP PROCESSES OF SAMPLE QUANTILES

被引:3
|
作者
FALK, M [1 ]
机构
[1] KATHOLISCHE UNIV EICHSTATT,FAK MATH GEOG,W-8078 EICHSTATT,GERMANY
关键词
BOOTSTRAP ESTIMATE; QUANTILE FUNCTION; SAMPLE QUANTILE; SMOOTHED BOOTSTRAP; KERNEL ESTIMATE; CONFIDENCE INTERVAL; FUNCTIONAL CENTRAL LIMIT THEOREM; BROWNIAN MOTION;
D O I
10.1016/0167-7152(91)90119-C
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
It is shown that the accuracy of the bootstrap estimate of the quantile function pertaining to the distribution of the sample q-quantile based on n independent and identically distributed observations is exactly O(P)(1/n1/4), q epsilon (0, 1) fixed. This rate can be improved considerably by applying smoothed bootstrap estimates. Our results are formulated in terms of functional central limit theorems for the corresponding inverse bootstrap processes.
引用
收藏
页码:529 / 536
页数:8
相关论文
共 50 条