ABSTRACT
The previous chapters of this book have been primarily concerned with the problem of extracting information from a given set of data provided by measurement sensors for accurate determination of parameter values which may have some physical significance. This gave rise to the use of various ‘alphabetical’ criteria related to the highest probability density region for the parameters. Nevertheless, other experiment design criteria are also possible and in order to demonstrate this issue, we devote the present chapter to a study of the design of the experimental conditions so that the experiment is maximally informative with respect to response prediction and the effectiveness of tests for model structure determination.
