[R-sig-ME] zero inflated poisson output
Ben Bolker
bbolker at gmail.com
Fri Jul 27 16:30:21 CEST 2012
Brickhill, Daisy <r01db11 at ...> writes:
> I am using the glmm.admb package to fit a zero-inflated poisson
> mixed model. What I would like to know is why when I use
> summary(model) the output only gives (what appears to be) the count
> process result. I.e. the binomial process, that deals with the
> zero/non zero is not shown. If you use zeroinfl() from library(pscl)
> it gives both 'Count model coefficients (poisson with log link)' AND
> 'Zero-inflation model coefficients (binomial with logit link)'. The
> glmm.admb output seems to give only the count model coefficients. I
> am using glmm.admb rather than zeroinfl() because I have a random
> effect. Can anyone tell me whether is a way I could extract the
> zero-inflation model coefficients from a glmm.admb model?
One thing to note is that the current version of glmmADMB
only fits a constant model for the zero-inflation process: if you
are interested in more complex zero-inflation models, you should
ask here again.
What version of glmmADMB are you using?
(i.e. what is the output of sessionInfo() ? )
With the most recent version, when I run this
example ...
## simple simulated zero-inflated Poisson example
### simulate values
set.seed(101)
d <- data.frame(f=factor(rep(LETTERS[1:10],each=10)),x=runif(100))
u <- rnorm(10,sd=2)
d$eta <- with(d,u[f]+1+4*x)
pz <- 0.3
zi <- rbinom(100,size=1,prob=pz)
d$y <- ifelse(zi,0,rpois(100,lambda=exp(d$eta)))
## fit
zipmodel <- glmmadmb(y~x+(1|f),data=d,family="poisson",zeroInflation=TRUE)
I do get information about the zero-inflation parameter estimates ...
At the moment glmmADMB is not building on r-forge, but I
am trying to fix it ...
Ben Bolker
More information about the R-sig-mixed-models
mailing list