Linear Toeplitz covariance structure models with optimal estimators of variance components Original Research Article

This paper deals with the problem of optimal quadratic unbiased estimation for statistical models with linear Toeplitz covariance structure. The main result is a necessary and sufficient condition for these models to have an optimal unbiased estimator for any linear combination of variance components. This result is obtained by means of special Jordan algebras which are a powerful tool to characterize optimality in quadratic unbiased estimation.