A few years back, I compared the pseudo random number generator in Splus
(version 3.0)
to a number of others (8 in all) and found no significant difference in looking
a 20 distributions
and sample sizes of n=10, 20, 40, 80, and 1000.
I was motivated to do so by a fellow who said "it doesn't match SAS exactly so
it must be wrong."
I have since learned the value of such a comment.
So I used the PRNGs in Splus to do my research, and have ever since.
Much ado about nothing, in my humble opinion.
Nonetheless, the American Statistician article is a worthwile exercise in
keeping everyone honest,
even if its not perfect.
Best wishes,
Eric Gibson, PhD
{the opinions expressed herein in no way represent those of Novartis
Pharmaceuticals}
"H. D. Vinod" <vinod@murray.fordham.edu> on 06/30/99 02:26:30 PM
To: s-news@wubios.wustl.edu
cc: (bcc: Eric Gibson/PH/Novartis)
Subject: [S] Does the new version of S-Plus fix num acc. and RNG problems?
Does the new version of S-PLUS solve some of the seriuos
numerical accuracy problems pointed out in McCullough (1999)?
For an academic user like me, too many new versions are not good.
We wish they will simply fix the older versions.
For those who may have missed The American Statistician article , you may be
interested to know that B. D. McCullough has assessed the numerical
reliability of S-PLUS, SAS, and SPSS. The articles can be downloaded from
the TAS website http://www.amstat.org/publications/tas/
It seems that the random number generator(RNG) for s-plus fails four of the
DIEHARD tests and
ANOVA fails SmnLsg9 problem
Nonlinear regression fails to get good R-square for the Lancsos1 problem.
H. D. Vinod
Professor of Economics, Fordham University, Bronx, NY 10458
vinod@murray.fordham.edu
Web page: www.fordham.edu/economics/vinod
-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news
-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news
|