Titlebook: Linear Statistical Inference; Proceedings of the I T. Caliński,W. Klonecki Conference proceedings 1985 Springer-Verlag Berlin Heidelberg 19

匍匐前進

Singular Information Matrices, Directional Derivatives, and Subgradients in Optimal Design Theory, Under the standard statistical assumptions the problem of finding an optimal approximate design leads to a convex minimization problem of the following type:.The optimality criterion φ is a convex and decreasing real-valued function on the set of all positive definite (s × s)-matrices, and J(M) is

daredevil

analogous

AVID

Some Remarks on Improving Unbiased Estimators by Multiplication with a Constant,ualifies to admissibility under the mean squared error risk function. Uniformly better estimates are easily given by multiplication of an unbiased estimator with some constant less than but close to one. The optimal choice of such a multiplier is investigated and more specific results are derived fo

過于平凡

On Improving Estimation in a Restricted Gauss-Markov Model,with respect to a weighted quadratic risk function. The first from the competitory estimators is based on the sample and linear restrictions contained in the model, while the second estimator additionally utilizes some linear restrictions which are not necessary fulfilled by the vector parameters. T

輕觸

Admissibility, Unbiasedness, and Nonnegativity in the Balanced, Random, One-Way Anova Model,ion. Pukelsheim (1976) noted that quadratics in the observations which are invariant to fixed effects follow a linear model in the variance components, and that results in linear estimation theory can then be applied directly to this derived linear model. However, this model has characteristics whic

BRINK

On Inference in a General Linear Model with an Incorrect Dispersion Matrix,investigated. Previously known results regarding the equality of best linear unbiased estimators and of test statistics under two linear models which differ in their dispersion matrices follow as special cases.

易發(fā)怒

A Split-Plot Design with Wholeplot Treatments in an Incomplete Block Design,re the number of wholeplots in each block may be less than the number of levels of factor A, is considered. The m levels of factor A are arranged in a proper incomplete block design. The h levels of factor B are arranged in a randomized complete block design within each level of factor A, by conside

Debark

Sensitivity of Linear Models with Respect to the Covariance Matrix, V. For this purpose we give conditions that the derivative of the GME with respect to a vanishes. In many cases it turns out that derivatives only can be zero in directions where the GME is also constant for some finite interval.

BRINK