A NOTE ON ROBUST DESIGNS FOR POLYNOMIAL REGRESSION

被引：16

作者：

DETTE, H ^{[1
]}

机构：

[1] UNIV GOTTINGEN,INST MATEMAT,W-3400 GOTTINGEN,GERMANY

来源：

JOURNAL OF STATISTICAL PLANNING AND INFERENCE | 1991年 / 28卷 / 02期

关键词：

POLYNOMIAL REGRESSION; ROBUST DESIGN; D-OPTIMAL DESIGN; CANONICAL MOMENTS; DISCRETE UNIFORM DISTRIBUTION;

D O I：

10.1016/0378-3758(91)90027-C

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 show that every discrete symmetric probability measure on [-1,1] with finite support can be characterized as a model robust D-optimal design (in the sense of Lauter (1974) in polynomial regression. As special cases the discrete uniform and some model robust designs given in the literature are identified as optimal designs for a class of polynomial regression models.

引用

页码：223 / 232

页数：10

共 22 条

[1] MODEL ROBUST, LINEAR-OPTIMAL DESIGNS [J].