Admissible linear estimators of an arbitrary vector of parametric functions in the general Gauss-Markov model (Q753353)

Some results about admissible linear estimation are obtained under the general Gauss-Markov model \(\{\) Y, \(X\beta\), \(\sigma\) \({}^ 2V\}\), with matrix X and V possibly deficient in rank. The author extends the admissible linear estimation, relaxing the assumption that the vector \(K\beta\) is identifiable, with K an \(m\times n\) real matrix. A general characterization is derived, of the class of admissible estimators for \(K\beta\) and \(\beta\). Finally, the admissibility of several estimators of \(\beta\) proposed in the literature is examined.