×

zbMATH — the first resource for mathematics

Sufficiency and completeness in the general Gauss-Markov model. (English) Zbl 0535.62007
Author’s summary: This paper develops the concepts ”linear sufficiency”, ”linear minimal sufficiency”, and ”linear completeness” in the general Gauss-Markov model \(EY=X\beta\), Cov Y\(=V\). Characterizations of these concepts are given. It is shown that linear sufficient and linear complete statistics are sufficient and complete, respectively, if Y is normally distributed. An application considers stepwise regression analysis.
Reviewer: C.Hipp

MSC:
62B05 Sufficient statistics and fields
62J05 Linear regression; mixed models
PDF BibTeX XML Cite