Density estimation using bootstrap bandwidth selector

被引：6
|
作者：
Bose, Arup [1 ]
Dutta, Santanu [2 ]
机构：
[1] Indian Stat Inst, Stat Math Unit, Kolktata 700108, India
[2] Tezpur Univ Napaam, Dept Math Sci, Tezpur 784028, Assam, India
来源：
STATISTICS & PROBABILITY LETTERS | 2013年 / 83卷 / 01期
关键词：
Kernel density estimator; Bootstrap; Plug-in; Cross-validation; Automatic bandwidth; INTEGRATED SQUARED ERROR; CHOICE;
D O I：
10.1016/j.spl.2012.08.027
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：
Smoothing methods for density estimators struggle when the shape of the reference density differs markedly from the actual density. We propose a bootstrap bandwidth selector where no reference distribution is used. It performs reliably in difficult cases and asymptotically outperforms well known automatic bandwidths. (C) 2012 Elsevier B.V. All rights reserved.
引用
收藏
页码：245 / 256
页数：12
相关论文
共 50 条
12345