[ADMB Users] restricted maximum likelihood

[Ian Fiske]

Thanks for the thoughtful responses.  I like your empiricism over
theology approach -- wish that was how statistics was taught!

Best,
Ian

On Fri, Apr 16, 2010 at 2:34 PM, dave fournier <otter at otter-rsch.com> wrote:
>
> I used the definition in Pinheiro and Bates on page 75.
>
> http://books.google.ca/books?id=N3WeyHFbHLQC&printsec=frontcover&dq=pinheiro+and+bates&source=bl&ots=pRaMVJuNrd&sig=s3-AX_s0PWDb1ecX2mwdBU0unMo&hl=en&ei=GBfJS9ndKoOusgPw-7H1BA&sa=X&oi=book_result&ct=result&resnum=4&ved=0CA8Q6AEwAw
>
> Interestingly I have seen Bates opine that  reml makes no sense for
> nonlinear mixed effects models. But that usually means that he just
> can't get something to work with inferior techniques (That is the usual
> R experts' approach to anything that can't be done in R).
> Link is https://stat.ethz.ch/pipermail/r-sig-mixed-models/2009q1/002104.html
> Personally I like to implement the estimation procedure and then
> evaluate its performance on simulated and real data rather than entering
> into theological discussions.
>
> _______________________________________________
> Users mailing list
> Users at admb-project.org
> http://lists.admb-project.org/mailman/listinfo/users
>



-- 
Ian Fiske
PhD Candidate
Department of Statistics
North Carolina State University