A Note on Optimal Bounded Designs
An equivalence theorem is formulated to characterize optimal designs with general prespecified direct constraints on the design intensity. As an application locally optimal designs are obtained with bounded intensity for the logistic regression model. Moreover, it is shown that, for additive models, optimal marginally bounded designs can be generated from their optimal counterparts in the corresponding marginal models.
Keywordsconstrained design equivalence theorem optimal design logistic regression bounded design intensity additive model marginal design
Unable to display preview. Download preview PDF.
- Dette, H. and Sahm, M. (1997). E-optimal designs for the double exponential model. In MODA 5 — Advances in Model-Oriented Data Analysis and Experimental Design, Eds A.C. Atkinson, L. Pronzato and H.R Wynn, pp. 11–20. Heidelberg: Physica-Verlag.Google Scholar
- Fedorov, V.V (1972). Theory of Optimal Experiments. New York: Academic Press.Google Scholar
- Müller, W.G. (1998). Collecting Spatial Data. Heidelberg: Physica-Verlag.Google Scholar
- Wynn, H.R (1982). Optimum submeasures with applications to finite population sampling. In Statistical Decision Theory and Related Topics III. Proc. Third Purdue Symp. Vol. 2., Eds S.S. Gupta and J.O. Berger, pp. 485–495. New York: Academic Press.Google Scholar