A Comparison of Marginal Likelihood Computation Methods

Charles S. BOS1
Faculty of Economics and Operations Research
Vrije Universiteit Amsterdam

February 20th, 2002

Abstract

In a Bayesian analysis, different models can be compared on the basis of the expected or marginal likelihood they attain. Many methods have been devised to compute the marginal likelihood, but simplicity is not the strongest point of most methods. At the same time, the precision of methods is often questionable.

In this paper several methods are presented in a common framework. The explanation of the differences is followed by an application, in which the precision of the methods is tested on a simple regression model where a comparison with analytical results is possible.

JEL classification: C11, C52, C63
Keywords: Marginal likelihood, Bayesian analysis

Footnotes:

1 Correspondence to Charles S. Bos, Faculty of Economics and Operations Research, Vrije Universiteit Amsterdam, De Boelelaan 1105, NL-1081 HV  Amsterdam, The Netherlands. Email: cbos@feweb.vu.nl, URL: http://www.tinbergen.nl/~cbos/. Numerous discussions with Herman K. van Dijk, Richard Paap and others are gratefully acknowledged.

[Main page] [Top] [Next: Software]
Last change: 5/3/2002