[R-sig-ME] Details control options for glmmADMB

[Ben Bolker]

ADMB runs its optimization in 'phases' where a progressively larger
subset of the parameters are estimated (in earlier phases, some
parameters are fixed to default values).  If you look in the definition
(TPL) file

file.show(system.file("tpl","glmmadmb.tpl",package="glmmADMB"))

and search for 'phase' you'll see that different phases are used for the
overdispersion parameters, zero-inflation parameters, random effects
standard deviations, random effects correlations.

  That said, I haven't actually seen a context where I've actually
needed to use this feature -- often you could just turn off the relevant
part of the model (e.g. zero-inflation) rather than setting
the maximum phase to a smaller value.

For more about phases, the ADMB manuals are at

http://admb-project.org/documentation/manuals/admb-user-manuals

specifically, the current version is at

http://ftp.admb-project.org/admb-11.4/manuals/admb-11.4.1.pdf

On 16-05-16 07:09 AM, Fraixedas, Sara wrote:
> Dear all,
> 
> I am a PhD student at the University of Helsinki and I would like to
> know what exactly the argument "maximum phase" means in the control
> options for the glmmADMB package. How does it help/contribute to
> solve convergence problems?
> 
> Best wishes,
> 
> 
> Sara Fraixedas Doctoral Student The Helsinki Lab of Ornithology
> (HelLO) Finnish Museum of Natural History P.O. Box 17 00014
> University of Helsinki, Finland Tel. +358-9-19128851
> 
> [[alternative HTML version deleted]]
> 
> _______________________________________________ 
> R-sig-mixed-models at r-project.org mailing list 
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>