[R-sig-ME] Zero-inflation with glmmadmb
Pin chanratana
chanratana.pin at gmail.com
Thu Dec 1 08:36:00 CET 2016
Hi,
I have some question regarding to zero-inflation and hurdle model
- I fit a glmmadmb model of a species camera-trapped at waterholes during
the dry season. The data contain 51.9% of zeros and the
variance/mean=10.41. Following is a fit model and its result:
m1 <- glmmadmb(WS ~ depth + offset(log(TN)), ndata4, "nbinom1")
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) -3.677 0.214 -17.16 <2e-16 ***
depth 0.318 0.136 2.34 0.02 *
Negative binomial dispersion parameter: 6.6339 (std. err.: 1.7732)
- And then I check for overdispersion
E <- resid(m1, type= 'pearson')
N <- nrow(ndata4)
p <- length(coef(m1))
sum(E^2)/(N-p) = 14.22
My questions are:
1. With the information above, should I go for more extend such as
zero-inflation and hurdle model?
2. To fit zero-inflation model, which approach is better way between
glmmadmb, zero-inflation and hurdle function?
Thanks,
[[alternative HTML version deleted]]
More information about the R-sig-mixed-models
mailing list