Rcpp/CppBugs example

[ttriche]

The existing (and well circulated) example of inlining a CppBugs run into R via Rcpp is

http://www.mail-archive.com/[email protected]/msg00825.html

An updated version with eg. LDA, a Dirichlet process mixture model, or an Indian Buffet Process would be nice.

The latter might be especially compelling as, to my knowledge, nothing like it exists in the R world, the closest comparable being the PyIBP project, which implements Stephen Walker and Finale Doshi-Velez's slice sampler, augmented to allow infinite real-valued features as suggested by Zoubin Ghahramani and others. Implementing the sampler via CppBugs seems like a logical direction inasmuch as npBayes methods are particularly sensitive to performance, and slice sampling was one of the first MCMC methods to be successfully sped up via GPU computation. But, that's just my $0.02.