This paper proposes a class of optimal design criteria for prediction in linear regression models with r responses based on the volume of the rectangular confidence region, termed RLr-optimality. A general equivalence theorem and a geometrical characterization of the RLr-optimal design are established, which are used to obtain or verify the RLr-optimality. Several examples are given for illustration.
All Science Journal Classification (ASJC) codes
- Statistics and Probability
- Statistics, Probability and Uncertainty
- Applied Mathematics