Bayesian prediction with approximate frequentist validity

We characterize priors which asymptotically match the posterior coverage probability of a Bayesian prediction region with the corresponding frequentist coverage probability. This is done considering both posterior quantiles and highest predictive density regions with reference to a future observation. The resulting priors are shown to be invariant under reparameterization. The role of Jeffreys' prior in this regard is also investigated. It is further shown that, for any given prior, it may be possible to choose an interval whose Bayesian predictive and frequentist coverage probabilities are asymptotically matched.

[1]  Jayanta K. Ghosh,et al.  Characterization of priors under which Bayesian and frequentist Barlett corrections are equivalent in the multiparameter case , 1991 .

[2]  R. Tibshirani.  Noninformative priors for one parameter of many , 1989 .

[3]  B. L. Welch,et al.  On Formulae for Confidence Points Based on Integrals of Weighted Likelihoods , 1963 .

[4]  David R. Cox,et al.  Prediction and asymptotics , 1996 .

[5]  J. Ghosh,et al.  Frequentist validity of highest posterior density regions in the multiparameter case , 1993 .

[6]  M. Ghosh,et al.  Second-order probability matching priors , 1997 .

[7]  Thomas J. DiCiccio,et al.  Frequentist and Bayesian Bartlett Correction of Test Statistics Based on Adjusted Profile Likelihoods , 1994 .

[8]  Trevor J. Sweeting,et al.  A framework for Bayesian and likelihood approximations in statistics , 1995 .

[9]  Paolo Vidoni.  A note on modified estimative prediction limits and distributions , 1998 .

[10]  F. Komaki.  On asymptotic properties of predictive distributions , 1996 .

[11]  Thomas A. Severini,et al.  On the Relationship between Bayesian and Non-Bayesian Interval Estimates , 1991 .

[12]  Seymour Geisser,et al.  8. Predictive Inference: An Introduction , 1995 .

[13]  D. Dey,et al.  Frequentist validity of posterior quantiles in the presence of a nuisance parameter : higher order asymptotics , 1993 .

[14]  Jayanta K. Ghosh,et al.  On priors providing frequentist validity for Bayesian inference , 1995 .

[15]  Thomas A. Severini,et al.  Bayesian interval estimates which are also confidence intervals , 1993 .

[16]  G. Datta.  On priors providing frequentist validity of Bayesian inference for multiple parametric functions , 1996 .

[17]  H. Jeffreys,et al.  The Theory of Probability , 1896 .

[18]  H. Kuboki.  Reference priors for prediction , 1998 .

[19]  Peter J. Bickel,et al.  A Decomposition for the Likelihood Ratio Statistic and the Bartlett Correction--A Bayesian Argument , 1990 .

[20]  Charles Stein,et al.  On the coverage probability of confidence sets based on a prior distribution , 1985 .

[21]  H. W. Peers.  On Confidence Points and Bayesian Probability Points in the Case of Several Parameters , 1965 .

[22]  Keying Ye,et al.  Frequentist validity of posterior quantiles for a two-parameter exponential family , 1996 .

[23]  Malay Ghosh,et al.  ON THE INVARIANCE OF NONINFORMATIVE PRIORS , 1996 .

[24]  Richard A. Johnson.  Asymptotic Expansions Associated with Posterior Distributions , 1970 .

[25]  John Aitchison,et al.  Statistical Prediction Analysis , 1975 .

[26]  Trevor J. Sweeting,et al.  On the construction of Bayes–confidence regions , 1999 .

[27]  Anna Nicolaou.  Bayesian intervals with good frequency behaviour in the presence of nuisance parameters , 1993 .