On Thu, 15 Jul 2010, Ana Navarrete, Anthropologisches Inst. wrote:
Dear Sir,
When requesting the summary of pglm analyses in R, I encounter repeatedly the
following warning message:
"Warning message:
In pf(Fstat, object$k - 1, object$n - object$k, ncp = 0, lower.tail = FALSE,
:
full precision was not achieved in 'pnbeta'"
This message appears both with the pglm estimate of Lambda and with Lambda
set to 0.
Could you tell me what should I correct to solve this problem?
This is a warning that you tried to evaluate a tail probability (I
guess a p-value) in an rather extreme case. What case, you didn't
deign to tell us (see the footer of this and every R-help message).
Very likely the result is good enough for your purposes, but only you
can decide that.
However, from what we do see, pf is called with ncp = 0. That is
almost certainly an error by the author of the code you are using (in
some package you didn't mention?), since (for consistency with values
of ncp near zero) pf(ncp=0) does not use the same algorithm as pf()
without supplying ncp.
I did try to track this down, but suspect you are not using any CRAN
nor BioC package, maybe one of a few possibilities on R-forge.
Best regards,
Ana Navarrete
*****************************************************************
Ana Navarrete
Anthropological Institute and Museum
Winterthurerstrasse 190
8057 Zurich
Tel: +41-798693918
E-mail: a.navarr...@aim.uzh.ch
______________________________________________
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.
--
Brian D. Ripley, rip...@stats.ox.ac.uk
Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/
University of Oxford, Tel: +44 1865 272861 (self)
1 South Parks Road, +44 1865 272866 (PA)
Oxford OX1 3TG, UK Fax: +44 1865 272595
______________________________________________
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.
