by Schwabe, R., Graßhoff, U., Großmann, H., Holling, H..
Series: 2005-38, Preprints
In this paper we show that product type designs are optimal in partially heteroscedastic multi-factor linear models. This result is applied to obtain locally D-optimal designs in multi-factor generalized linear models by means of a canonical transformation. As a consequence we can construct optimal designs for direct logistic response as well as for Bradley-Terry type paired comparison experiments.
optimal design, logistic model, generalized linear model, heteroscedastic linear model, additive model, product type design, paired comparison, Bradley-Terry model