A couple of people have expressed an interest in having a look at the
problematic
code/package which led my cri de coeur posted yesterday.
I have therefore made the package accessible by putting it up on my
UNB web page:
http://www.math.unb.ca/~rolf
Click on ``Hidden generalized linear Markov model package.''
The package includes a file ``scr.test'' in the ``inst'' directory.
(It shows up in the
hglmm directory in your library when you install the package.)
In the script ``UA'' and ``CA'' are set equal to false. Hence the
nlm() based procedure runs,
giving a sensible answer and one warning message about ``Inf/NA''.
If you set UA to be TRUE (and leave CA FALSE) the procedure runs but
converges to a
silly answer and gives very many warning messages.
If you set both of these to be TRUE the procedure falls over
immediately, saying that there
is a coding problem with the ``analytic'' hessian.
Thanks to anyone who is willing to take a look at the package and
perhaps give me a hint
as to what is going wrong when the analytic hessian is used.
cheers,
Rolf
######################################################################
Attention:\ This e-mail message is privileged and confidenti...{{dropped}}
______________________________________________
R-help@r-project.org mailing list
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.
