Designs for generalized linear models (Q2799896)

This long, lucidly written and very informative review article has the flavor of a research monograph. It begins with an introduction to generalized linear models and touches upon the special cases of normal, binomial, Poisson distribution, and gamma data. Then, the theory of optimal experimental design is introduced, with a discussion on D-optimal designs, design efficiency and the problem of parameter uncertainty. Locally optimal designs are next taken up, and illustrated with logistic regression in one variable. This is followed by a discussion on optimal designs for two or more explanatory variables in the framework of binomial, gamma and Poisson data. Finally, designs with dependent data are discussed. The article concludes with a discussion on extensions and further reading. The extensive list of references will certainly endear it to its readers.NEWLINENEWLINEFor the entire collection see [Zbl 1327.62001].