[R-sig-ME] mixed zero-inflated poisson regression

Mollie Brooks mbrooks at ufl.edu
Tue Nov 15 22:57:54 CET 2011


Hi Ivar,
I have recently needed to do several regressions of zero inflated Poisson and negative binomial data where the covariates affect both the conditional mean and the zero inflation and there are random effects on the conditional mean. I've been using AD Model Builder and R2admb. I could send you documented example code if you want to try ADMB. You'd first have to get ADMB working on your computer. It gives the maximum likelihood values and uses Laplace approximation or GHQ on the random effects. Then you can do MCMC sampling starting from the mle if you need a better picture of the parameters' distributions.
cheers,
Mollie

Mollie Brooks
Ph.D. Candidate
NSF IGERT Fellow
Biology Department
University of Florida
mbrooks at ufl.edu
http://people.biology.ufl.edu/mbrooks




More information about the R-sig-mixed-models mailing list