Activity Number:
|
353
|
Type:
|
Contributed
|
Date/Time:
|
Wednesday, August 14, 2002 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Section on Statistics & the Environment*
|
Abstract - #301457 |
Title:
|
D-optimal Designs for the Poisson Regression Model in Impaired Reproduction Studies
|
Author(s):
|
Yanping Wang*+ and Eric Smith and Keying Ye
|
Affiliation(s):
|
Virginia Polytechnic Institute and State University and Virginia Polytechnic Institute and State University and Virginia Polytechnic Institute and State University
|
Address:
|
Statistics Department, Virginia Tech, Blacksburg, Virginia, 24060, U.S.A.
|
Keywords:
|
D-optimal Design ; Poisson Exponential Model
|
Abstract:
|
Optimal designs for generalized linear models have received increasing attention in recent years. Yet most of the current work focuses on the one variable first-order logistic regression model. In this paper, we extend this topic to the Poisson regression model. For the first-order model, optimal designs in terms of effective dose (ED) levels are independent of the parameters. However, for more complex models this is not the case. We investigated how D-optimal designs depend on the parameters for the one-variable quadratic model and the two-variable interaction model. We also studied the performance of some "standard" designs which might be intuitive to practitioners.
|