Benchmark Priors for Bayesian Model Averaging

In contrast to a posterior analysis given a particular sampling model, posterior model probabilities in the context of model uncertainty are typically rather sensitive to the specification of the prior. In particular, “diffuse” priors on model-specific parameters can lead to quite unexpected consequences. Here we focus on the practically relevant situation where we need to entertain a (large) number of sampling models and we have (or wish to use) little or no subjective prior information. We aim at providing an “automatic” or “benchmark” prior structure that can be used in such cases. We focus on the Normal linear regression model with uncertainty in the choice of regressors. We propose a partly noninformative prior structure related to a Natural Conjugate g-prior specification, where the amount of subjective information requested from the user is limited to the choice of a single scalar hyperparameter g0j . The consequences of different choices for g0j are examined. We investigate theoretical properties, such as consistency of the implied Bayesian procedure. Links with classical information criteria are provided. More importantly, we examine the finite sample implications of several choices of g0j in a simulation study. The use of the MC3 algorithm of Madigan and York (1995), combined with efficient coding in Fortran, makes it feasible to conduct large simulations. In addition to posterior criteria, we shall also compare the predictive performance of different priors. A classic example concerning the economics of crime will also be provided and contrasted with results in the literature. The main findings of the paper will lead us to propose a “benchmark” prior specification in a linear regression context with model uncertainty.

[1]  Isaac Ehrlich,et al.  The Deterrent Effect of Capital Punishment: A Question of Life and Death , 1973 .

[2]  J. Richard.  Posterior and Predictive Densities for Simultaneous Equation Models , 1973 .

[3]  I. Ehrlich.  Participation in Illegitimate Activities: A Theoretical and Empirical Investigation , 1973, Journal of Political Economy.

[4]  G. Schwarz.  Estimating the Dimension of a Model , 1978 .

[5]  Edward E. Leamer,et al.  Specification Searches: Ad Hoc Inference with Nonexperimental Data , 1980 .

[6]  J. Bernardo.  Expected Information as Expected Utility , 1979 .

[7]  B. G. Quinn,et al.  The determination of the order of an autoregression , 1979 .

[8]  S. Geisser,et al.  A Predictive Approach to Model Selection , 1979 .

[9]  J. Bernardo.  A Bayesian analysis of classical hypothesis testing , 1980 .

[10]  D. Spiegelhalter,et al.  Bayes Factors and Choice Criteria for Linear Models , 1980 .

[11]  A. Zellner,et al.  Posterior odds ratios for selected regression hypotheses , 1980 .

[12]  Charles A. Ingene,et al.  Specification Searches: Ad Hoc Inference with Nonexperimental Data , 1980 .

[13]  George E. P. Box,et al.  Sampling and Bayes' inference in scientific modelling and robustness , 1980 .

[14]  A. Atkinson.  Likelihood ratios, posterior odds and information criteria , 1981 .

[15]  H. Akaike.  Likelihood of a model and information criteria , 1981 .

[16]  G. Chow.  A comparison of the information and posterior probability criteria for model selection , 1981 .

[17]  L. R. Pericchi,et al.  An alternative to the standard Bayesian procedure for discrimination between normal linear models , 1984 .

[18]  Mark F. J. Steel,et al.  Bayesian-analysis of Systems of Seemingly Unrelated Regression Equations Under a Recursive Extended Natural Conjugate Prior Density , 1988 .

[19]  T. J. Mitchell,et al.  Bayesian Variable Selection in Linear Regression , 1988 .

[20]  D. Poirier.  Frequentist and Subjectivist Perspectives on the Problems of Model Building in Economics , 1988 .

[21]  L. Bauwens.  The "pathology" of the natural conjugate prior density in the regression model , 1990 .

[22]  Arnold Zellner,et al.  Bayesian and non-Bayesian methods for combining models and forecasts with applications to forecasting international growth rates , 1993 .

[23]  E. George,et al.  Journal of the American Statistical Association is currently published by American Statistical Association. , 2007 .

[24]  S. Chib,et al.  Regression models under competing covariance matrices: A Bayesian perspective , 1993 .

[25]  Christopher Cornwell,et al.  Estimating the Economic Model of Crime with Panel Data , 1994 .

[26]  J. Geweke,et al.  Variable selection and model comparison in regression , 1994 .

[27]  A. Gelfand,et al.  Bayesian Model Choice: Asymptotics and Exact Calculations , 1994 .

[28]  D. Madigan,et al.  Model Selection and Accounting for Model Uncertainty in Graphical Models Using Occam's Window , 1994 .

[29]  Dean P. Foster,et al.  The risk inflation criterion for multiple regression , 1994 .

[30]  Adrian E. Raftery,et al.  Simultaneous Variable and Transformation Selection in Linear Regression , 1995 .

[31]  Adrian E. Raftery,et al.  Accounting for Model Uncertainty in Survival Analysis Improves Predictive Performance , 1995 .

[32]  J. York,et al.  Bayesian Graphical Models for Discrete Data , 1995 .

[33]  L. Wasserman,et al.  A Reference Bayesian Test for Nested Hypotheses and its Relationship to the Schwarz Criterion , 1995 .

[34]  David Draper,et al.  Assessment and Propagation of Model Uncertainty , 2011 .

[35]  Peter C. B. Phillips,et al.  Bayesian model selection and prediction with empirical applications , 1995 .

[36]  D. Madigan,et al.  Eliciting prior information to enhance the predictive performance of Bayesian graphical models , 1995 .

[37]  Purushottam W. Laud,et al.  Predictive Model Selection , 1995 .

[38]  A. O'Hagan,et al.  Fractional Bayes factors for model comparison , 1995 .

[39]  P. Green.  Reversible jump Markov chain Monte Carlo computation and Bayesian model determination , 1995 .

[40]  S. Chib,et al.  Understanding the Metropolis-Hastings Algorithm , 1995 .

[41]  Hugh Chipman,et al.  Bayesian variable selection with related predictors , 1995, bayes-an/9510001.

[42]  J. Berger,et al.  The Intrinsic Bayes Factor for Model Selection and Prediction , 1996 .

[43]  M. Clyde,et al.  Prediction via Orthogonalized Model Mixing , 1996 .

[44]  R. Kohn,et al.  Nonparametric regression using Bayesian variable selection , 1996 .

[45]  A. Raftery.  Approximate Bayes factors and accounting for model uncertainty in generalised linear models , 1996 .

[46]  D. Madigan,et al.  A method for simultaneous variable selection and outlier identification in linear regression , 1996 .

[47]  Joseph G. Ibrahim,et al.  Predictive specification of prior model probabilities in variable selection , 1996 .

[48]  Dean Phillips Foster,et al.  Calibration and Empirical Bayes Variable Selection , 1997 .

[49]  D. Madigan,et al.  Bayesian Model Averaging in Proportional Hazard Models: Assessing the Risk of a Stroke , 1997 .

[50]  D. Madigan,et al.  Bayesian Model Averaging for Linear Regression Models , 1997 .

[51]  E. George,et al.  APPROACHES FOR BAYESIAN VARIABLE SELECTION , 1997 .

[52]  J. Lindsey,et al.  Some statistical heresies , 1999 .