Summary
In this paper we continue the study of the extension of the Gauss-Markov theorem to certain general kinds of multiresponse models. In particular we obtain necessary and sufficient conditions, for the general incomplete multiresponse (GIM) model and the multiple design multi-response (MDM) model, such that unique best linear unbiased estimates (BLUE's) exist for all elements in a subset of the set of all estimable linear functions of the location parameters. Also the theory is illustrated by a couple of nontrivial examples.
Similar content being viewed by others
References
Ogawa, J. and Ishii, G. (1965). The relationship algebra and the analysis of variance of a PBIB design,Ann. Math. Statist.,36, 1815–1828.
Roy, S. N. and Srivastava, J. N. (1964). Hierarchical, andp-block multiresponse designs and their analysis,Sankhya, Mahalanobis Volume, 419–428.
Srivastava, J. N. (1967). On the extension of Gauss-Markov theorem to complex multivariate linear models,Ann. Inst. Statist. Math.,19, 417–437.
Srivastava, J. N. (1968)., On a general class of designs for multiresponse experiments,Ann. Math. Statist.,39, 1825–1843.
Trawinski, I. M. (1961). Incomplete-variable designs (unpublished thesis), V.P.I., Blacksburg, Virginia.
Trawinski, I. M. and Bargmann, R. E. (1964). Maximum likelihood estimation with incomplete multivariate data,Ann. Math. Statist.,35, 647–657.
Author information
Authors and Affiliations
Additional information
This research was wholly supported by the U.S. Air Force under grant No. F33615-67-C-1436, monitored by the Aerospace Research Laboratories.
About this article
Cite this article
Srivastava, J.N., McDonald, L.L. On the extensions of Gauss-Markov theorem to subsets of the parameter space under complex multivariate linear models. Ann Inst Stat Math 25, 383–393 (1973). https://doi.org/10.1007/BF02479384
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF02479384