Local improvement of best linear unbiased estimation and admissibility under the weakly singular gauss-markov model

Abstract

Under the weakly singular Gauss-Markov model, the class of linearly admissible estimators for the expectation of the observable random vector with respect to the mean square error criterion is considered. It is demonstrated that this class admits linearly admissible estimators for an arbitrary estimable parametric function, which locally improve the best linear estimator with respect to the mean square error matrix criterion.

Keywords:

• general ridge estimator
• linear estimator
• lowner partial ordering
• mean square error (matrix)