[R-sig-ME] Profile likelihood using R2admb

Stirrup, Oliver oliver.stirrup.13 at ucl.ac.uk
Fri Apr 25 12:24:00 CEST 2014

Dear Prof Bolker

Thank you for the quick response, and for developing this package in the first place!

Best wishes


Oliver Stirrup
PhD Student
MRC Clinical Trials Unit at UCL

Message: 2
Date: Thu, 24 Apr 2014 07:39:24 -0400
From: Ben Bolker <bbolker at gmail.com>
To: r-sig-mixed-models at r-project.org
Subject: Re: [R-sig-ME] Profile likelihood using R2admb
Message-ID: <5358F7EC.3090203 at gmail.com>
Content-Type: text/plain; charset=ISO-8859-1

On 14-04-24 06:01 AM, Stirrup, Oliver wrote:
> Hello all,
> I am currently experimenting with various mixed model structures using 
> the R2admb package. Everything is working smoothly, except that I have 
> been unable to make use of the functionality relating to likelihood 
> profiles.
> If I attempt to run 'do_admb' with the argument 'profile=TRUE', I 
> receive the error message 'must specify profpars when 
> checkparam=='write' and profile is TRUE'. However, if I then try to 
> also specify the parameters to include, for example adding the 
> argument 'profpars=c("rho")', I receive the error message 'unused 
> argument (profpars=c("rho"))'.
> I am using R version 3.1.0 and R2admb version 0.7.10, and I have 
> noticed that the R2admb manual states that the profile option for 
> do_admb is 'untested!'. Has anyone else encountered this problem? If 
> so, is there any way around it?

  I (the package maintainer) haven't tried this myself in a while, and when I did I found it didn't work!  I imagine this is a fairly minor glitch and will try to fix it soon:
https://github.com/bbolker/R2admb/issues/13 -- you may end up having to install the github version (via devtools::install_github()) once I have it fixed

  Ben Bolker

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