Consistency of Bayesian Estimates for the Sum of Squared Normal Means with a Normal Prior

We consider the problem of estimating the sum of squared means when the data (x 1,...,x n ) are independent values with x i ∼ N(θ i , 1) and θ 1, θ 2... are a priori i.i.d. N(0, σ 2) with σ 2 known. This example has posed difficulties for many approaches to inference. We examine the consistency properties of several estimators derived from Bayesian considerations. We prove that a particular Bayesian estimate (LRSE) is consistent in a wider set of circumstances than other Bayesian estimates like the posterior mean and mode. We show that the LRSE is either equal to the positive part of the UMVUE or differs from it with a relative error no greater than 2/n. We also prove a consistency result for interval estimation and discuss checking for prior-data conflict. While it can be argued that the choice of the N(0,σ 2) prior is inappropriate when σ 2 is chosen large to reflect noninformativity, this argument is not applicable when σ 2 is chosen to reflect knowledge about the unknowns. As such it is important to show that there are consistent Bayesian estimation procedures using this prior.

[1]  Charles Stein,et al.  An Example of Wide Discrepancy Between Fiducial and Confidence Intervals , 1959 .

[2]  Milton Abramowitz,et al.  Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables , 1964 .

[3]  M. Stone,et al.  Marginalization Paradoxes in Bayesian and Structural Inference , 1973 .

[4]  Gun Ho Jang,et al.  Inferences from prior-based loss functions , 2011, 1104.3258.

[5]  Further remarks on estimating the parameter of a noncentral chi-square distribution , 1976 .

[6]  Michael Evans,et al.  A limit result for the prior predictive applied to checking for prior-data conflict , 2011 .

[7]  Mo Suk Chow A complete class theorem for estimating a noncentrality parameter , 1987 .

[8]  Michael Evans,et al.  Weak Informativity and the Information in One Prior Relative to Another , 2011, 1201.1766.

[9]  K. L. Saxena,et al.  Estimation of the Non-Centrality Parameter of a Chi Squared Distribution , 1982 .

[10]  A. Gelfand Estimation in noncentral distributions , 1983 .

[11]  James O. Berger,et al.  ESTIMATION OF QUADRATIC FUNCTIONS: NONINFORMATIVE PRIORS FOR NON-CENTRALITY PARAMETERS , 1998 .

[12]  Roger L. Berger Comment on Perlman and Wu, “The Emperor’s new tests” (with rejoinder by authors) , 1999 .

[13]  D. Waal Bayes estimate of the noncentrality parameter in multivariate analysis , 1974 .

[14]  Christian P. Robert,et al.  Estimation of noncentrality parameters , 1993 .

[15]  Masaaki Sibuya,et al.  Generalized Hypergeometric Distributions , 2006 .

[16]  R. Wolpert,et al.  Integrated likelihood methods for eliminating nuisance parameters , 1999 .

[17]  Michael Evans,et al.  Checking for prior-data conflict , 2006 .

[18]  Michael Evans,et al.  Hypothesis Assessment and Inequalities for Bayes Factors and Relative Belief Ratios , 2013 .

[19]  Tim B. Swartz,et al.  Optimally and computations for relative surprise inferences , 2006 .

[20]  Michael Evans,et al.  Bayesian ikference procedures derived via the concept of relative surprise , 1997 .

[21]  M. Evans,et al.  Optimal properties of some Bayesian inferences , 2007, 0710.1151.

[22]  M. Perlman,et al.  Some remarks on estimating a noncentrality parameter , 1975 .