Haff, L. R. - In: Journal of Multivariate Analysis 6 (1976) 2, pp. 265-280
Empirical Bayes estimators are given for the mean of a k-dimensional normal distribution, k = 3. We assume that y ~ Nk([theta], V1), V1 = diag(vi), vi known (i = 1, 2,..., k); also, [theta] ~ Nk(0, V2) - V2 defined by one or more unknown parameters. Of particular interest is V2 generated by an...