Bayesian model selection offers a coherent framework for identifying the most plausible models when the number of candidate predictors greatly exceeds the number of observations. Central to this ...