ON MULTIVARIATE KERNEL ESTIMATION FOR SAMPLES FROM WEIGHTED DISTRIBUTIONS

被引：39

作者：

AHMAD, IA ^{[1
]}

机构：

[1] NO ILLINOIS UNIV,DIV STAT,DE KALB,IL 60115

来源：

STATISTICS & PROBABILITY LETTERS | 1995年 / 22卷 / 02期

关键词：

D O I：

10.1016/0167-7152(94)00057-F

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：

Data from multivariate weighted distributions appear in such cases as missing data, damaged data, sociological or economic data. Multivariate kernel density estimation is discussed for these data where its mean square error (pointwise and integrated) is derived and for an important and common special case, its mean absolute error is outlined. The analogous problem dealing with regression estimation is presented.

引用

页码：121 / 129

页数：9