On Bayesian Model Assessment and Choice Using Cross-Validation Predictive Densities: Appendix

Short description of the prior and the MCMC specification details is given here. See (Neal, 1996, 1997, 1999; Lampinen & Vehtari, 2001) and the FBM software manual (Neal, 2000) for additional details. The MCMC sampling was done with the FBM1 software and Matlab-code derived from the Netlab2 toolbox. FBM was used when possible (normal and logistic models) because of its speed. Matlab-based implementation was slower, but it was much easier to implement new features to it. In the following, we use the notation r ∼ F(a) as shorthand for p(r) = F(r |a) where a denotes the parameters of the distribution F , and the random variable argument r is not shown explicitly. N (μ, σ 2) denotes a normal distribution with mean μ and variance σ 2. Parametrization of the Inverse Gamma here is