Re: [R] error in kmeans

From: Gavin Simpson <gavin.simpson_at_ucl.ac.uk>
Date: Sat 02 Apr 2005 - 02:46:59 EST

Asha Jayanthi wrote:
> I am trying to generate kmean of 10 clusters for a 165 x 165 matrix.
>
> i do not see any errors known to me. But I get this error on running the
> script
>
> Error: empty cluster: try a better set of initial centers
>
> the commands are
>
> M <-matrix(scan("R_mutual",n = 165 * 165),165,165,byrow = T)
>
> cl <- kmeans(M,centers=10,20)
> len = dim(M)[1]
> ....
> ....
>
> I ran the same script last night and it was working prefectly. I have
> not made any changes at all !!!And this is very strange. This evening
> when I ran the same script i am getting this error. My matrix file is
> also untouched.
>
> Can any one let me know how to go about this. I must generate 10-mean
> clusters
> Is there anyother way of doing it ? and how to avoid such error in future?
>
> Asha
>

Reading ?kmeans we have:

   centers: Either the number of clusters or a set of initial cluster

           centers. If the first, a random set of (distinct) rows in 'x'
           is chosen  as the initial centers.

So each time you run your analysis kmeans will select 10 random starting values for the cluster centers. Sometimes the selection ends up with no objects in a cluster, sometimes it doesn't - it is random (pseudo) after all. You could provide the centers yourself of course, something along the lines of (adapted from Venables and Ripley (1999) Modern Applied Statistics with Splus, 3rd Edition page 338 - not sure about 4th Ed as my copy is at home just now):

M <- data.frame(matrix(rnorm(5000), ncol = 25)) M.x <- as.matrix(M)
h <- hclust(dist(M.x), method = "average") initial <- tapply(M.x, list(rep(cutree(h, 10),

                                 ncol(M.x)),
                             col(M.x)),
                             mean)

M.km <- kmeans(M.x, initial)

HTH

-- 
%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%
Gavin Simpson                     [T] +44 (0)20 7679 5522
ENSIS Research Fellow             [F] +44 (0)20 7679 7565
ENSIS Ltd. & ECRC                 [E] gavin.simpsonATNOSPAMucl.ac.uk
UCL Department of Geography       [W] http://www.ucl.ac.uk/~ucfagls/cv/
26 Bedford Way                    [W] http://www.ucl.ac.uk/~ucfagls/
London.  WC1H 0AP.
%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%

______________________________________________
R-help@stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Received on Sat Apr 02 05:10:51 2005

This archive was generated by hypermail 2.1.8 : Fri 03 Mar 2006 - 03:30:59 EST