[R] help on nlm (gradient) (fwd)

Luz Maria Palacios Derflingher palacios at math.ucalgary.ca
Mon Jun 19 16:28:16 CEST 2006


No worries anymore. Figured it out.

Thanks for everything.


---------- Forwarded message ----------
Date: Sun, 18 Jun 2006 23:12:25 -0600 (MDT)
From: Luz Maria Palacios Derflingher <palacios at math.ucalgary.ca>
To: r-help at lists.R-project.org
Cc: r-help at stat.math.ethz.ch.
Subject: help on nlm (gradient)


I am having some trouble using nlm in R for windows version 2.2.1. I have to 
minimise a function which includes in its insides the evaluation of 
integrals through the "adapt" function.
In order to make R work less, I calculated the gradient and coded it. I made 
sure that the code and function for the gradient worked.
So I know that the main function (the one to minimise,let's call it "func1") 
yields values when evaluated.
I know that the gradient I coded also works and yields values when evaluated 
(on its own); when I put it all together and use nlm I get the following 

Error in nlm(func1, starting values,  :
         probable coding error in analytic gradient

Any light that can be shed on this would be highly appreciated.
Many thanks

Luz Palacios
PhD Candidate University of Calgary

More information about the R-help mailing list