[R] problem with cclust[er] package

[Igor Oleinik]

Here is a bit of outer loop too:

			library(cclust)
			# clustering algorithm using k-means
			bDone <- 0
			ncntrs <- 2
			browser()
			while (bDone == 0) {
			
				# k-means with ncntrs (changes from 1 to
number of points) centers
				# problem starts here on the second call in
debugging mode
				# res1 is 16X52 matrix

				cl <- cclust(res1, ncntrs, iter.max  = 20,
verbose = TRUE, dist="manhattan", method="kmeans")

				# max distance from center to to any point
in cluster
				meandist <- sqrt(max(cl$withinss/cl$size)) *
20/52

				if (meandist < 0.1) bDone = 1
				
				ncntrs <- ncntrs + 1
				if (ncntrs > seascnt) bDone = 1

			} # while (bDone = 0)


I am running Win2000, R 1.6.2. I installed R
over the Internet.

Igor.



-----Original Message-----
From: Thomas W Blackwell [mailto:tblackw at umich.edu]
Sent: Wednesday, March 05, 2003 3:31 PM
To: Igor Oleinik
Subject: RE: [R] problem with cclust[er] package 



I think you will need to give us some of the outer loop,
the one which calls cclust() repeatedly and tests for the
"certain conditions".  I would look for the problem there,
not inside cclust itself ... but I could be wrong.  Also
what system, version, method of installation  ?

-  tom blackwell  -  u michigan medical school  -  ann arbor  -

('Fraid I'm offline until tomorrow.)




On Wed, 5 Mar 2003, Igor Oleinik wrote:

> I have checked that section already.
> Sorry, I should have mentioned that.
>
> Memory limit increase does not work.
> Installtion of msvcrt.dll does not work
> either.
>
> Thank you.
>
> -----Original Message-----
> From: ripley at stats.ox.ac.uk [mailto:ripley at stats.ox.ac.uk]
> Sent: Wednesday, March 05, 2003 2:44 PM
> To: Igor Oleinik
> Cc: r-help at stat.math.ethz.ch
> Subject: Re: [R] problem with cclust[er] package
>
>
> On Wed, 5 Mar 2003, Igor Oleinik wrote:
>
> > I am calling cclust function in cclust package
> > repeatedly until some ceratain conditions
> > for a cluster are met. Unfortunately,
> > the system crashes on the second call (after debugging).
> >
> > # kmeans res1 is a well defined matrix
> > cl <- cclust(res1, as.numeric(ncntrs), iter.max  = 20, verbose = FALSE,
> > dist="manhattan", method="kmeans")
> >
> >
> > RGui has generated errors and will be closed by Windows. ...
> >
> > What might be the problem?
>
> There is a section in the rw-FAQ on how to find out.
>
> --
> Brian D. Ripley,                  ripley at 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 at stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-help
>