Key points are not available for this paper at this time.
We review the across‐model simulation approach to computation for Bayesian model determination, based on the reversible jump Markov chain Monte Carlo method. Advantages, difficulties and variations of the methods are discussed. We also discuss some limitations of the ideal Bayesian view of the model determination problem, for which no computational methods can provide a cure.
Hastie et al. (Wed,) studied this question.