MV-optimization in Simple Linear Regression
MV-optimality is a potentially difficult criterion because of its nondifferentiability at equal variance designs. However in many cases such designs can be easily determined. In this paper MV-optimum designs for simple linear regression are found. The equivalence theorem of and the directional derivative of the MV-criterion derived by Ford I., (3), have been used for this purpose. It turns out that for simple linear regression there exist an MV-optimal design with a support of at most two points. Such designs could be of a wide ranging practical value.
KeywordsPoint Design Simple Linear Regression Information Matrix Directional Derivative Equivalence Theorem
Unable to display preview. Download preview PDF.
- DETTE H. and STUDDEN W.J. (1993). Minimax Designs in Linear Regression Models. Preprint.Google Scholar
- ELFVING G. (1959). Design of Linear Experiments. In Probability and Statistics. The Harald Cramér Volume ( U. Grenander, ed.). Stockholm: Almagrist, and Wiksell, pp 58–74.Google Scholar
- FORD I. (1976). Ph.D. Thesis. University of Glasgow.Google Scholar
- LOPEZ-FIDALGO J. (1992). Minimizing the Largest of the Parameters Variances. V(ß)-optimality. Proceedings of the 3rd International Workshop MODA 3 in Petrodvorets, Russia. Physica-Verlag. New York.Google Scholar
- TAKEUCHI K. (1961). On the Optimality of Cetain Type of PBIB designs. Rep. Stat. Appl. Res. Un. Japan Sci. Engrs., Vol. 8 (3), pp 140–145.Google Scholar
- TORSNEY B. and ALAHMADI, A.M. (1995). Designing for Minimal Dependent Observations. Statistica Sinica (to appear).Google Scholar