Hi Michael, thanks for your answer. Indeed, with a 100x100 matrix it runs even pretty fast with k=30. But as with a lot of things in R, there is a disproportionate rise in the calculation time once you exceed a certain size limit on your matrices. In the end, it ran about 8 hours for my complete matrix.
Thanks for the suggestion, that saves me quite a bit of time. For the full story: I'm applying a new distance measure for comparing phylogenetic trees. Now the space this is calculated in, has to be mapped back on an euclidean space, which I'm trying out right now. I noticed that the 2D-solution seems a bad representation of the real distances, so I increase the dimensionality. I use the dimensions to get a medoid and a centroid in the euclidean space, but those results obviously depend on the number of dimensions used in the MDS. So I'm trying to figure out when these location measures get more or less stable. graphical representation is off course bound to 2 dimensions. 3D plots tend to be confusing with over 800 points. Tried color coding, but that doesn't really help... Cheers Joris On Wed, May 26, 2010 at 5:32 AM, Michael Denslow <michael.dens...@gmail.com>wrote: > Hi Joris, > > On Tue, May 25, 2010 at 1:00 PM, Joris Meys <jorism...@gmail.com> wrote: > > Dear all, > > > > I'm running a set of nonparametric MDS analyses, using a wrapper for > isoMDS, > > on a 800x800 distance matrix. I noticed that setting the parameter k to > > larger numbers seriously increases the calculation time. Actually, with > k=10 > > it calculates already longer than for k=2 and k=5 together. It's now > > calculating for 6 hours, and counting... > > Seems like a long time, I have a 100x100 matrix that takes about 40 > secs to run with k=10. What is the wrapper function doing? > > > > > There is quite a difference between the results using k=2 or k=5 when > > looking at the first 2 dimensions (logically...). I suspect the same when > > k=10. Yet, I start asking myself whether this makes sense if I'm only > using > > the first 2 dimensions. And I can't think of a formal method to check in > a > > nMDS framework how much dimensions are enough. Anybody an idea? > > You might want to look at the nmds.min() function in the ecodist > package, which seeks to minimize stress. Out of curiosity, do you > often use 10 dimensional solutions in your field of study? > > Hope this helps, > Michael > > > I use metaMDS from the vegan package, although it's not really meant to > be > > used on these data. > > > > Cheers > > Joris > > > > -- > > Joris Meys > > Statistical Consultant > > > > Ghent University > > Faculty of Bioscience Engineering > > Department of Applied mathematics, biometrics and process control > > > > Coupure Links 653 > > B-9000 Gent > > > > tel : +32 9 264 59 87 > > joris.m...@ugent.be > > ------------------------------- > > Disclaimer : http://helpdesk.ugent.be/e-maildisclaimer.php > > > > [[alternative HTML version deleted]] > > > > ______________________________________________ > > 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. > > > > > > -- > Michael Denslow > > I.W. Carpenter Jr. Herbarium [BOON] > Department of Biology > Appalachian State University > Boone, North Carolina U.S.A. > -- AND -- > Communications Manager > Southeast Regional Network of Expertise and Collections > sernec.org > > 36.214177, -81.681480 +/- 3103 meters > -- Joris Meys Statistical Consultant Ghent University Faculty of Bioscience Engineering Department of Applied mathematics, biometrics and process control Coupure Links 653 B-9000 Gent tel : +32 9 264 59 87 joris.m...@ugent.be ------------------------------- Disclaimer : http://helpdesk.ugent.be/e-maildisclaimer.php [[alternative HTML version deleted]] ______________________________________________ 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.
