Forward search added-variable t-tests and the effect of masked outliers on model selection

被引：33

作者：

Atkinson, AC ^{[1
]}

Riani, M

机构：

[1] Univ London London Sch Econ & Polit Sci, Dept Stat, London WC2A 2AE, England

[2] Univ Parma, Dipartimento Econ, I-43100 Parma, Italy

来源：

BIOMETRIKA | 2002年 / 89卷 / 04期

关键词：

Box-Cox transformation; projection; score statistic; t-distribution; transformation to normality; very robust methods;

D O I：

10.1093/biomet/89.4.939

中图分类号：

Q [生物科学];

学科分类号：

07 [理学]; 0710 [生物学]; 09 [农学];

摘要：

Monitoring the t-tests for individual regression coefficients in 'forward' search fails to identify the importance of observations to the significance of the individual regressors. This failure is due to the ordering of the data by the search. We introduce an added-variable test which has the desired properties since the projection leading to residuals destroys the effect of the ordering. An example illustrates the effect of several masked outliers on model selection. Comments are given on the related test for response transformations.

引用

页码：939 / 946

页数：8

共 11 条

[1]

[Anonymous], ROBUST DIAGNOSTIC RE

[2]

ATKINSON AC, 1973, J ROY STAT SOC B MET, V35, P473

[3]

ATKINSON AC, 1989, BIOMETRIKA, V76, P223

[4]

Tests in the fan plot for robust, diagnostic transformations in regression [J].

Atkinson, AC ;

Riani, M .

CHEMOMETRICS AND INTELLIGENT LABORATORY SYSTEMS, 2002, 60 (1-2) :87-100

[5]

Atkinson AC., 1985, Plots, transformations and regression

[6]

an introduction to graphical methods of diagnostic regression analysis

[7]

Breiman L, 1996, ANN STAT, V24, P2350

[8]

Cook R. D., 1982, RESIDUALS INFLUENCE

[9]

A method for simultaneous variable selection and outlier identification in linear regression [J].

Hoeting, J ;

Raftery, AE ;

Madigan, D .

COMPUTATIONAL STATISTICS & DATA ANALYSIS, 1996, 22 (03) :251-270

[10]

Neter J., 1990, APPL LINEAR STAT MOD

← 1 2 →