ABSTRACT

The calculation of the posterior probabilities for a set of competing models is essential in the Bayesian approach for model selection. However, the marginal likelihood is generally not easily computed.