[R-sig-ME] estimating evidence: practical bridge / importance sampling code?

[Ryan King]

Hi all, I'm doing MCMCglmm estimation of a mixed model and nested
submodel and would like to compute the bayes factor.  It looks like
using (warp) bridge sampling is the current favorite.  I see

1.Overstall, A.M. & Forster, J.J. Default Bayesian model determination
methods for generalised linear mixed models. Computational Statistics
& Data Analysis 54, 3269-3288 (2010).
2.Sinharay, S. & Stern, H.S. An Empirical Comparison of Methods for
Computing Bayes Factors in Generalized Linear Mixed Models. Journal of
Computational and Graphical Statistics 14, 415-435 (2005).
3.Ardia, D., Hoogerheide, L.F. & Dijk, H.K.V. To Bridge, to Warp or to
Wrap? A Comperative Study of Monte Carlo Methods for Efficient
Evaluation of Marginal Likelihood. SSRN eLibrary at
<http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1349876>

of which only 1. seems to have code.  Does anyone have practical
experience with this for GLMMs and recommendations?  I'm curious if
there are any implementations which use the more aggressive warps of
Meng and Schilling (reflections, addressing skewness) instead of just
mean-(co)variance / mode-curvature.

Thanks,
C Ryan King
Dept Health Studies
University of Chicago