Skip to content
Prev 74 / 20628 Next

Extraordinary accuracy of mcmcsamp?

On 2/22/07, T.C. Cameron <T.C.Cameron at leeds.ac.uk> wrote:
The MCMC chain is stuck, which can happen with generalized linear
mixed models.  The Metropolis-Hastings algorithm that I had concocted
for MCMC sampling from GLMMs worked well on the examples that I tried
but does not work as well as I had hoped in general.

You will need to formulate your own sampler or use the z statistics.