Well-known shortcomings, advantages and computational challenges in Bayesian modelling: a few case stories

Bayesian inference can be used to judge the data fit quantitatively through the marginal likelihood. In many practical cases only one model is considered and parameter averaging is simply used to avoi