Robust normal reference bandwidth for kernel density estimation

被引：10

作者：

Zhang, Jin ^{[1
]}

Wang, Xueren ^{[1
]}

机构：

[1] Yunnan Univ, Sch Math & Stat, Kunming 650091, Yunnan, Peoples R China

来源：

STATISTICA NEERLANDICA | 2009年 / 63卷 / 01期

基金：

加拿大自然科学与工程研究理事会; 中国国家自然科学基金;

关键词：

biased cross-validation; least squares cross-validation; mean integrated squared error; outliers; CROSS-VALIDATION;

D O I：

10.1111/j.1467-9574.2008.00392.x

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：

Bandwidth selection is the main problem of kernel density estimation, the most popular method of density estimation. The classical normal reference bandwidth usually oversmoothes the density estimate. The existing hi-tech bandwidths have computational problems (even may not exist) and are not robust against outliers in the sample. A highly robust normal reference bandwidth is proposed, which adapts to different types of densities.

引用

页码：13 / 23

页数：11

共 17 条

[1]

BOWMAN AW, 1984, BIOMETRIKA, V71, P353

[2] DENSITY-ESTIMATION AND SUICIDE RISKS IN PSYCHIATRIC-TREATMENT [J].

COPAS, JB ;

FRYER, MJ .

JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES A-STATISTICS IN SOCIETY, 1980, 143 :167-176

[3]

HALL P, 1991, J ROY STAT SOC B, V53, P245

[4]

Hardle W., 1991, SMOOTHING TECHNIQUES

[5] EXACT MEAN INTEGRATED SQUARED ERROR [J].

MARRON, JS ;

WAND, MP .

ANNALS OF STATISTICS, 1992, 20 (02) :712-736

[6] ESTIMATION OF A PROBABILITY DENSITY-FUNCTION AND MODE [J].

PARZEN, E .

ANNALS OF MATHEMATICAL STATISTICS, 1962, 33 (03) :1065-&

[7] REMARKS ON SOME NONPARAMETRIC ESTIMATES OF A DENSITY-FUNCTION [J].

ROSENBLATT, M .

ANNALS OF MATHEMATICAL STATISTICS, 1956, 27 (03) :832-837

[8]

RUDEMO M, 1982, SCAND J STAT, V9, P65

[9] BIASED AND UNBIASED CROSS-VALIDATION IN DENSITY-ESTIMATION [J].

SCOTT, DW ;

TERRELL, GR .

JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 1987, 82 (400) :1131-1146

[10]

Scott W.D., 1992, Multivariate Density Estimation

← 1 2 →