by Soumaya, M.; Schwabe, R..
Series: 2011-27, Preprints
In applications often more than one dependent variable is observed in each experimental unit. In some of these situations the explanatory variables may be adjusted separately for the components in these models. For example, if one is interested in both pharmacokinetics and pharmacodynamics, the time points need not be identical for the measurements of the two quantities within one subject. As the observations will be correlated within one unit, the data may be described by a multivariate model, which has the structure of a seemingly unrelated regression.
multivariate linear model, seemingly unrelated regression, D-optimal design, product type design