A comparison of design alternatives for generalized linear models

被引:0
作者
Arnoldsson, G [1 ]
机构
[1] Mid Sweden Univ, Dept Informat, S-85170 Sundsvall, Sweden
关键词
generalized linear models; binary response; experimental design; D-optimality; allocation; efficiency;
D O I
10.1080/03610929808832653
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
Two-point or three-point designs for estimation in two-parameter generalized linear models are compared with respect to D-efficiency. The models are restricted to four binary response models. Results indicate that a non-optimal design may still perform well compared to the globally D-optimal design.
引用
收藏
页码:107 / 120
页数:14
相关论文
共 11 条
[1]   EXPERIMENTAL-DESIGN FOR CATEGORIZED DATA [J].
ABDELBASIT, KM ;
PLACKETT, RL .
INTERNATIONAL STATISTICAL REVIEW, 1981, 49 (02) :111-126
[2]  
[Anonymous], OPTIMAL DESIGN EXPT
[3]   SEQUENCES CONVERGING TO D-OPTIMAL DESIGNS OF EXPERIMENTS [J].
ATWOOD, CL .
ANNALS OF STATISTICS, 1973, 1 (02) :342-352
[4]   A NOTE ON OPTIMAL BAYESIAN DESIGN FOR NONLINEAR PROBLEMS [J].
CHALONER, K .
JOURNAL OF STATISTICAL PLANNING AND INFERENCE, 1993, 37 (02) :229-235
[5]  
FORD I, 1992, J ROY STAT SOC B MET, V54, P569
[6]  
JAIN RK, 1987, STAT HEFTE, V28, P285
[7]  
McCullagh P., 1989, GEN LINEAR MODELS, DOI [DOI 10.1007/978-1-4899-3242-6, 10.1201/9780203753736, DOI 10.2307/2347392]
[8]   OPTIMAL DESIGNS FOR BINARY DATA [J].
MINKIN, S .
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 1987, 82 (400) :1098-1103
[9]  
Silvey S.D, 1980, OPTIMAL DESIGN, V1, DOI 10.1007/978-94-009-5912-5
[10]  
SITTER RR, 1993, SCAND J STAT, V20, P329