[R] clustering question ... hclust & kmeans

[Mark Robinson]

> Have you tried this?  Hierarchical clustering on 3200 items takes quite a
> lot of memory, so I hope you have lots.

Windows doesn't like it much but the 512Mb RAM helps.

> cutree will cut a tree into k(=400) clusters, and return a vector of group
> membership just like kmeans.

Many Thanks.


I also want to be able to do this 400-cluster approach in kmeans.  I am
getting a message like the following, for even 40 clusters and 2000
iterations.

> xxx<-kmeans( gene.clus, 40, 2000 )
Error in switch(Z$ifault, stop("empty cluster: try a better set of initial
centers"),  :
        empty cluster: try a better set of initial centers

Perhaps doing kmeans in this way is a unreasonable.  I plan to make the
initial clusters as the 400 most representative clusters from "cutree"
above.  Maybe that'll help

I would also like to explore divisive clustering as well.  From my
understanding, this is not included in the R "cluster" library.  Anybody
written anything to do this?

M.

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._