If you walk through the code line by line, you can probably find what
you want, then make relatively simple changes to the code so these
things are exported. However, it may not be easy to use, because you
need to understand the parameterization. To obtain that, you may need
to study the code more carefully. It may also help to study Pinheiro
and Bates (2000) Mixed-Effects Models in S and S-PLUS (Springer) and
some of Bates' more recent publications cited in the help files in the
nlme and lme4 packages.
What problem are you trying to solve? Hessian matrices are most
useful for normal approximations, which are however sometimes not very
good. Better approximations are obtained from profile likelihood, and
even better ones are available from Monte Carlo. If you'd like more
help from this list, you might increase your chances of getting a quick,
useful reply by considering the suggestions in the posting guide,
"www.R-project.org/posting-guide.html".
hope this helps.
spencer graves
Ingmar Visser wrote:> I use the lme function from the nlme library (or alternatively from the
> Matrix library) to estimate a random effects model. Both functions return
> the covariance matrix of the estimated parameters. I have the following
> question:
> Is it possible to retrieve the information matrix of such a model (ie from
> the fitted object)? In particular, the information matrix can be computed
as
> a sum of individual contributions:
> Sum_i H_i = H,
> where H_i are the contributions of each case to the Hessian matrix H.
> Are these individual contributions computed somewhere in either of the
> functions to estimate random effects models, and is it possible to extract
> them somehow from the fitted objects?
>
> ingmar
--
Spencer Graves, PhD
Senior Development Engineer
PDF Solutions, Inc.
333 West San Carlos Street Suite 700
San Jose, CA 95110, USA
spencer.graves at pdf.com
www.pdf.com <http://www.pdf.com>
Tel: 408-938-4420
Fax: 408-280-7915