[R-sig-ME] Has anyone tried Whit Armstrong's CppMCMC library?

[Whit Armstrong]

Thanks for the the post, Professor Bates.

CppBugs can be found here: http://github.com/armstrtw/CppBugs

It's a combination of WinBUGS and PyMC.

The api is still very unstable as I'm attempting to navigate between
good design and maintaining familiar WinBUGS conventions.

Rcpp is the best way to use it from R.  I'll add an example of calling
it this way on the front page so that everyone can try it.

My thinking is that since it's already very inconvenient to run a BUGS
model from R (writing the BUGS script, providing inits, passing data
back and forth), it's not really asking too much for users to write a
small c++ class inline in an R script that is the CppBugs equivalent
of a BUGS model.

I'm very keen to have help on this project, so if anyone is interested
please contact me.

I'll send a follow up post highlighting a few of the features of
CppBugs and examples of WinBUGS vs PyMC conventions that I've followed
in the design.

-Whit




On Tue, Nov 2, 2010 at 11:08 AM, Douglas Bates <bates at stat.wisc.edu> wrote:
> Whit Armstrong has written a C++ library called CppMCMC that provides
> BUGS-like functionality in compiled code.  I see that one of his
> examples uses the contagious bovine pleuropneumonia data from the cbpp
> data set in the lme4 package, fitting an overdispersed binomial-like
> model to it.
>
> Does anyone have experience with these classes?  This is not directly
> an R question as, at present, I don't know of a bridge between CppMCMC
> and R (although it would be an interesting project to use Rcpp for
> this).
>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>