On a Two-Stage Estimate of the Mean

Abstract

A two-stage estimate for the mean vector of a multivariate distribution is proposed for the situation in which the prior knowledge about the mean vector can be expressed in the form of an initial estimate. The proposed estimate is compared with the sample mean based on a single sample of equivalent size. The proposed estimate is also compared with the Bayes' estimate based on an equivalent sample. The latter comparison has been accomplished only for the univariate case.