Criterion-Robust Optimal Designs for Fourier Regression Models

• Activity Number: 47
• Type: Contributed
• Date/Time: Sunday, August 11, 2002 : 4:00 PM to 5:50 PM
• Sponsor: General Methodology
• Author(s): Mei-Mei Zen*+ and Min-Hsiao Tsai
• Affiliation(s): National Cheng-kung University and National Cheng-kung University
• Keywords: optimal designs ; selection criterion ; Fourier regression ; multiple-objective

Abstract

Consider the problem of discriminating between the regression models of degree k-1 (Model A) and k (Model B) on an interval [a, b]. The goal here is to find experiment designs that can efficiently achieve all three objectives simultaneously: (1) to discriminate between Model A and Model B, and depending on the decision, (2) to make inferences in either Model A, or (3) Model B. For the dual problem, objective (1) is one which we must pay attention to, and either objective (2) or (3) is the other one. In Pukelsheim and Rosenberger (1993), they presented several designs for the above consideration for polynomial regression models with k=3.. Zen and Tsai (2001) proposed a multiple-objective selection criterion (namely, Mr-criterion), which put various weights for different objectives, and derived the corresponding weighted D-optimal designs for polynomial regression models. In this study, we consider the Fourier regression models on an interval . Under Mr-criterion, the weighted D-optimal designs will be derived in terms of canonical moments. The efficiency of the weighted D-optimal designs will be studied under various weight .