×

Optimal elimination of nuisance parameters in mixed linear models. (English) Zbl 0760.62067

Author’s abstract: The vector parameter of the mean value of an observation vector in a mixed linear model (MLM) is supposed to be divided into necessary and nuisance vector parameters. A class of linear transformations of the observation vector eliminating the nuisance vector parameter which do not cause a loss of information about the necessary parameter is considered. The problem which of these transformations has the property that the same locally best quadratic estimator of variance components from the original and the transformed MLM is obtained is solved. Several kinds of estimators are investigated.

MSC:

62J10 Analysis of variance and covariance (ANOVA)
62J05 Linear regression; mixed models
62F10 Point estimation
PDFBibTeX XMLCite
Full Text: EuDML

References:

[1] FELLMAN J.: Linear models with nuisance parameters - some theoretical results applied to balanced designs. Meddelanden working papers. Swedish School of Economics and Business Administrations. Working papers 147, 1986.
[2] FELLMAN J.: Estimation in linear models with nuisance parameters. Statistics & Decisions, Supplement issue No 2, 1985, 161-164. · Zbl 0609.62109
[3] KLEFFE J.: Simultaneous estimation of expectation and covariance matrix in linear models. Zentralinstitut für Mathematik und Mechanik der Akademie der Wissenschaften der DDR, Berlin 1977. · Zbl 0415.62026 · doi:10.1080/02331887808801444
[4] KUBÁČEK L., KUBÁČKOVÁ L.: Statistical aspect of eliminating systematic effects. Studia geoph. et geod. 30, 1986, 1-12.
[5] KUBÁČEK L.: Elimination of nuisance parameters in a regression model. Math. Slovaca 36, 1986, 137-144. · Zbl 0605.62081
[6] RAO C. R., MITRA S. K.: Generalized Inverse of Matrices and Its Application. N. York, J. Wiley 1971.
[7] RAO C. R., KLEFFE J.: Estimation of variance components. P. R. Krishnaiah (Ed.), Handbook of Statistics. Vol. I. North-Holland, Amsterdam 1980. · Zbl 0476.62058
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.