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 g[0j]. The consequences of different choices for g[0j] are examined. We investigate theoretical properties, such as consistency of the implied Bayesian procedure. Links with classical information criteria are provided. In addition, we examine the finite sample implications of several choices of g[0j] in a simulation study. The use of the MC^3 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.
(This abstract was borrowed from another version of this item.)
References listed on IDEAS
Please report citation or reference errors to , or , if you are the registered author of the cited work, log in to your RePEc Author Service profile, click on "citations" and make appropriate adjustments.:
- Akaike, Hirotugu, 1981. "Likelihood of a model and information criteria," Journal of Econometrics, Elsevier, vol. 16(1), pages 3-14, May.
- Atkinson, A. C., 1981. "Likelihood ratios, posterior odds and information criteria," Journal of Econometrics, Elsevier, vol. 16(1), pages 15-20, May.
- Hoeting, Jennifer & Raftery, Adrian E. & Madigan, David, 1996. "A method for simultaneous variable selection and outlier identification in linear regression," Computational Statistics & Data Analysis, Elsevier, vol. 22(3), pages 251-270, July.
- Phillips, Peter C. B., 1995.
"Bayesian model selection and prediction with empirical applications,"
Journal of Econometrics,
Elsevier, vol. 69(1), pages 289-331, September.
- Peter C.B. Phillips, 1992. "Bayesian Model Selection and Prediction with Empirical Applications," Cowles Foundation Discussion Papers 1023, Cowles Foundation for Research in Economics, Yale University.
- Chow, Gregory C., 1981. "A comparison of the information and posterior probability criteria for model selection," Journal of Econometrics, Elsevier, vol. 16(1), pages 21-33, May.
- Richard, J. F. & Steel, M. F. J., 1988. "Bayesian analysis of systems of seemingly unrelated regression equations under a recursive extended natural conjugate prior density," Journal of Econometrics, Elsevier, vol. 38(1-2), pages 7-37.
- Smith, M. & Kohn, R., .
"Nonparametric Regression using Bayesian Variable Selection,"
Statistics Working Paper
_009, Australian Graduate School of Management.
- Smith, Michael & Kohn, Robert, 1996. "Nonparametric regression using Bayesian variable selection," Journal of Econometrics, Elsevier, vol. 75(2), pages 317-343, December.
- Min, C.K. & Zellner, A., 1992.
""Bayesian and Non-Bayesian Methods for Combining Models and Forecasts with Applications to Forecasting International Growth Rates","
90-92-23, California Irvine - School of Social Sciences.
- Min, Chung-ki & Zellner, Arnold, 1993. "Bayesian and non-Bayesian methods for combining models and forecasts with applications to forecasting international growth rates," Journal of Econometrics, Elsevier, vol. 56(1-2), pages 89-118, March.
- repec:adr:anecst:y:1993:i:32:p:04 is not listed on IDEAS
- Bauwens, L., 1990.
"The "Pathology" Of The Natural Conjugate Prior Density In The Regression Model,"
90a14, Universite Aix-Marseille III.
- Luc Bauwens, 1991. "The 'pathologie' of the Natural Conjugate Prior Density in the Regression Model," Annals of Economics and Statistics, GENES, issue 23, pages 49-64.
- Isaac Ehrlich, 1973.
"The Deterrent Effect of Capital Punishment: A Question of Life and Death,"
NBER Working Papers
0018, National Bureau of Economic Research, Inc.
- Ehrlich, Isaac, 1975. "The Deterrent Effect of Capital Punishment: A Question of Life and Death," American Economic Review, American Economic Association, vol. 65(3), pages 397-417, June.
- Gary S. Becker, 1974.
"Crime and Punishment: An Economic Approach,"
in: Essays in the Economics of Crime and Punishment, pages 1-54
National Bureau of Economic Research, Inc.
- Poirier, Dale J, 1988. "Frequentist and Subjectivist Perspectives on the Problems of Model Building in Economics," Journal of Economic Perspectives, American Economic Association, vol. 2(1), pages 121-44, Winter.
- Chib, S. & Osiewalski, J. & Steel, M.F.J., 1990.
"Regression models under competing covariance matrices : A Bayesian perspective,"
1990-63, Tilburg University, Center for Economic Research.
- Chib, B. & Osiewalski, J. & Steel, M., 1990. "Regression Models Under Competing Covariance Matrices: A Baysian Perspective," Papers 9063, Tilburg - Center for Economic Research.
- Cornwell, Christopher & Trumbull, William N, 1994. "Estimating the Economic Model of Crime with Panel Data," The Review of Economics and Statistics, MIT Press, vol. 76(2), pages 360-66, May.
- Ehrlich, Isaac, 1973. "Participation in Illegitimate Activities: A Theoretical and Empirical Investigation," Journal of Political Economy, University of Chicago Press, vol. 81(3), pages 521-65, May-June.
When requesting a correction, please mention this item's handle: RePEc:fda:fdaddt:9806. See general information about how to correct material in RePEc.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: (Carmen Arias)
If references are entirely missing, you can add them using this form.