COMPROMISE DESIGNS IN HETEROSCEDASTIC LINEAR-MODELS

被引：11

作者：

DASGUPTA, A ^{[1
]}

MUKHOPADHYAY, S ^{[1
]}

STUDDEN, WJ ^{[1
]}

机构：

[1] PURDUE UNIV,DEPT STAT,W LAFAYETTE,IN 47907

来源：

JOURNAL OF STATISTICAL PLANNING AND INFERENCE | 1992年 / 32卷 / 03期

基金：

美国国家科学基金会;

关键词：

COMPROMISE DESIGNS; OPTIMAL DESIGNS; HETEROSCEDASTICITY; CHEBYSHEV POLYNOMIALS; CANONICAL MOMENTS; LINEAR AND POLYNOMIAL REGRESSIONS;

D O I：

10.1016/0378-3758(92)90017-M

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：

We consider heteroscedastic linear models in which the variance of a response is an exponential or a power function of its mean. Such models have earlier been considered in Bickel (1978), Carroll and Ruppert (1982) etc. Classical as well as Bayes optimal experimental design is considered. We specifically address the problem of 'compromise designs' where the experimenter is simultaneously interested in many estimation problems and wants to find a design that has an efficiency of at least 1/(1+epsilon) in each problem. For specific models we work out the smallest epsilon for which such a design exists. This is done for classical as well as Bayes problems. The effect of the variance function on the value of the smallest epsilon is examined. The maximin efficient design is then compared to the usual A-optimal design. Some general comments are made.

引用

页码：363 / 384

页数：22

共 41 条

[1] USING RESIDUALS ROBUSTLY .1. TESTS FOR HETEROSCEDASTICITY, NONLINEARITY [J].