A one-step robust estimator for regression based on the weighted likelihood reweighting scheme

被引：25

作者：

Agostinelli, C

Markatou, M

机构：

[1] Univ Padua, Dept Stat, I-35121 Padua, Italy

[2] Columbia Univ, Dept Stat, New York, NY 10027 USA

来源：

STATISTICS & PROBABILITY LETTERS | 1998年 / 37卷 / 04期

关键词：

breakdown point; efficiency; influence; robust; weighted likelihood;

D O I：

10.1016/S0167-7152(97)00136-3

中图分类号：

O21 [概率论与数理统计]; C8 [统计学];

学科分类号：

020208 [统计学]; 070103 [概率论与数理统计]; 0714 [统计学];

摘要：

We propose a one-step estimator for the vector of regression and error-scale parameters in a linear repression model. The estimator is asymptotically normal and fully efficient. Given appropriate initial values it achieves very low bias and high breakdown point. (C) 1998 Elsevier Science B.V. All rights reserved.

引用

页码：341 / 350

页数：10

共 24 条

[1]

[Anonymous], J COMPUTATIONAL GRAP

[2]

Brownlee K. A., 1965, STAT THEORY METHODOL

[3]

TRANSFORMATIONS IN REGRESSION - A ROBUST ANALYSIS [J].