The Loss of Efficiency Estimating Linear Functions under Restrictions

From MaRDI portal





This paper is focused on the problem of choosing \(X=\sum^{-1/2}X^{0}\) or \(X^{*}\) to estimate a linear function of the parameter components, in particular in ANOVA models. Here \(X^{0}\) is normally distributed with mean vector \(\theta^{0}\) and variance-covariance matrix \(\sigma^{2}\sum,\) where \(\sum\) is a known positive definite matrix and the vector \(X^{*}\) is the maximum likelihood estimator (MLE) of \(\theta,\) where \(\theta=\sum^{-1/2}\theta^{0}.\) It is proved that if the restrictions are given by a tree order or a simple order then the MLE of some contrast has greater mean square error (MSE) than the unrestricted estimator. A similar behaviour of the MLE is exhibited in a general restricted setting by a multivariate normal distribution with mean vector constrained to belong to a circular cone. The approach focuses on the central direction of the cones. These directions appear to have the greater MSE when the dimension of the restricted cone is big enough.











This page was built for publication: The Loss of Efficiency Estimating Linear Functions under Restrictions

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4956057)