[R-sig-ME] problems with glmmadmb
Ben Bolker
bbolker at gmail.com
Wed Sep 9 01:29:46 CEST 2015
No, not really.
* Gamma models are generally less well-tested. There might be
something wrong with the residuals calculation; I can take a look
(more likely to look if you send a reproducible example).
* As detailed in ?convergence in the lme4 documentation, the
convergence tests give a lot of false positives. If the estimated
coefficients & log-likelihoods match between glmmadmb and glmer I
think it's OK to trust them.
On Tue, Sep 8, 2015 at 3:19 AM, Daniel Harrison
<daniel at earthoceanspace.com> wrote:
> Hey there,
>
> I am having some problems with glmmadmb.
>
> I am predicting a series of glmm with gamma distribution.
>
> Glmmadmb converges on all models, however the pearson residuals from every
> model come back as all ŒNA¹.
>
> In glmmer several of the models fail to converge.
>
> Is this expected behaviour with a gamma distribution?
>
> Regards,
>
> Daniel Harrison
> PhD Candidate
> University of Sydney
>
>
>
> [[alternative HTML version deleted]]
>
>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>
More information about the R-sig-mixed-models
mailing list