On 27/04/20 3:01 am, J C Nash wrote:
:
Peter is correct. I was about to reply when I saw his post.
It should be possible to suppress the Hessian call. I try to do this
generally in my optimx package as computing the Hessian by finite differences
uses a lot more compute-time than solving the optimization problem that
precedes the usual Hessian calculation.
It may be time to consider some update to optim() in that regard, or to
put in some exception handling. It isn't likely in any main-line part
of optim() but in the post-solution phase of the routines. I'd welcome
some discussion on that with a view to improvement, but am not sure if
it should be R-devel or R-package-dev lists or off-list.
Thanks to Peter and John for this advice.
As to "suppressing the Hessian call" --- if I understand you correctly,
this is not a problem with optim(); in fact by default optim() does not
attempt to calculate the hessian. The problem arises with
MASS::fitdistr() which sets hessian=TRUE in the call to optim(),
willy-nilly.
cheers,
Rolf
--
Honorary Research Fellow
Department of Statistics
University of Auckland
Phone: +64-9-373-7599 ext. 88276
______________________________________________
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
