An embedded and charset-unspecified text was scrubbed... Name: not available Url: https://stat.ethz.ch/pipermail/r-help/attachments/20040118/e4ac60c7/attachment.pl
multcomp, simint, simtest and computation duration
3 messages · Patrick Giraudoux, Torsten Hothorn, Douglas Bates
Dear R-listers, I am trying to compute simultaneous confidence intervals with simint from the package multcomp. 230 measures (abundance) have been taken in 23 sites (factor) of a data.frame (donnees: a file can be sent on request, saved with save(donnees,file="donnees")). I would like to get all pairwise comparisons with : mc<- simint(ren~ID,type="Tukey",data=donnees)
you try to solve a (23^2 - 23) = 506 dimensional integration problem via some form of Monte-Carlo technique. If the sample sizes in are balanced, you can use the `TukeyHSD' function. Torsten
I cannot get a result in a reasonable time (after 2 hours the computer was still working and I interrupted the process). Can anybody tell me if there is some capacity limitation for simint (as well as for simtest)? Multicomp of Splus handles the problem adequately in a short time, but Splus is practically not accessible for student training... Kind regards, Patrick Giraudoux [[alternative HTML version deleted]]
______________________________________________ R-help at stat.math.ethz.ch mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Torsten Hothorn <Torsten.Hothorn at rzmail.uni-erlangen.de> writes:
I am trying to compute simultaneous confidence intervals with simint from the package multcomp. 230 measures (abundance) have been taken in 23 sites (factor) of a data.frame (donnees: a file can be sent on request, saved with save(donnees,file="donnees")). I would like to get all pairwise comparisons with : mc<- simint(ren~ID,type="Tukey",data=donnees)
you try to solve a (23^2 - 23) = 506 dimensional integration problem via some form of Monte-Carlo technique. If the sample sizes in are balanced, you can use the `TukeyHSD' function.
Even without balanced sample sizes you can use TukeyHSD, although it is most reliable if your sample sizes are close to balanced.