LINEAR ESTIMATORS OF THE MEAN VECTOR IN LINEAR MODELS:
PROBLEM OF ADMISSIBILITY
Abstract: In this paper we consider linear estimators in linear models with a general
covariance structure. Necessary conditions for admissibility of the linear estimators with
quadratic loss function are given and they are shown to be sufficient when only positive
definite covariance matrices are admitted. In the case where the set of admitted covariance
matrices coincides with all nonnegative definite matrices, it is shown that is admissible
for the expected value if and only if the eigenvalues of the matrix are in the closed
interval .
2000 AMS Mathematics Subject Classification: Primary: -; Secondary: -;
Key words and phrases: -