Bayesian variance estimation in the Gaussian sequence model with partial information on the means (Q2286367)

The authors of the paper under review consider the Gaussian sequence model under the assumption that a fixed fraction of the means is known, and study the problem of variance estimation from a frequentist Bayesian perspective (since the maximum likelihood estimator of variance is biased and inconsistent). They find that the marginal posterior is inconsistent for any i.i.d. prior on the mean parameters. However, it is demonstrated that consistency can be retained for a hierarchical prior based on Gaussian mixtures. In this case the authors establish a limiting shape result and determine the limit distribution (which is non-Gaussian). A numerical simulation study illustrates that the Bayesian analysis leads to new statistical estimators outperforming the correctly calibrated maximum likelihood estimator.