[ADMB Users] mcmc and -noinit -- use -mcmc2 instead
mbrooks at ufl.edu
Fri Mar 23 15:03:14 PDT 2012
Great, I got the hybrid mcmc example working with the following call
./hybridsimple_re -hybrid -hyeps .3 -hynsteps 20 -mcmc2 1000
I'll post this on the examples page soon.
So now I'll have the same model running once with regular mcmc, once with -mcrb 5, and once with -hybrid. I accidentally threw out results from -mcgrope. I wish I had time to do a thorough comparison of these methods, but this is not the focus of my current research. I'll probably go with hybrid since it seems to be well respected in the literature. Any users out there want to write a paper comparing these methods? Can I have the results before this chapter of my dissertation is due? just kidding.
On 23 Mar 2012, at 12:31 PM, dave fournier wrote:
> BTW in the hybrid mcmc beta regression mixed model example
> I did, the problematic term(s) were elements of the estimated covariance matrix for
> the RE's.
> Users mailing list
> Users at admb-project.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Users