[R] Problem with nlm() function.

bbolker bolker at ufl.edu
Mon Sep 17 17:36:24 CEST 2007




Rolf Turner-3 wrote:
> 
> 
> In the course of revising a paper I have had occasion to attempt to  
> maximize a rather
> complicated log likelihood using the function nlm().  This is at the  
> demand of a referee
> who claims that this will work better than my proposed use of a home- 
> grown implementation
> of the Levenberg-Marquardt algorithm.
> 
> I have run into serious hiccups in attempting to apply nlm().
> 
> [snip]
> 
> My gut feeling is that the problem is just a bit too complicated for  
> nlm() to handle, and
> I would simply leave it and go with the other approaches which appear  
> to work consistently
> and well.  However Referees Must Be Satisfied.  Has anyone any  
> suggestions as to
> how I can get nlm() to work with the analytical gradient and hessian?
> 

  This kind of problem is the perfect one to fall through the cracks on
R-help:
the poster has obviously done his homework (and thus cannot be
flamed to oblivion), a non-trivial problem that will take some work, not in
one of the particular areas of interest on someone on the list, and not
a problem where someone is going to be able to create an astoundingly
clever one-line solution that works 0.34 seconds faster than an intelligible
5-line solution.

  I don't think there's any conceivable way to solve this without access
to the aforementioned package.  Could you post it please?

  cheers
    Ben Bolker

-- 
View this message in context: http://www.nabble.com/Problem-with-nlm%28%29-function.-tf4463043.html#a12738419
Sent from the R help mailing list archive at Nabble.com.



More information about the R-help mailing list