[R] equivalent of the Splus function "eboulis()"

n.bouget@laposte.net n.bouget at laposte.net
Tue May 25 15:06:12 CEST 2004


This function already exists in the package multidim (see
http://cran.r-project.org/doc/packages/multidim.pdf)
What i want to do is to choose the good number of cluster
after using the function hclust on 6934 instances. 
> hc <- hclust(dist(AGRIINSTTableFinaleCR), "ward")
I use the following expression
> barplot(sort((hc$height/sum(hc$height))*100,decreasing=TRUE))
and i choose as number of cluster the number before a big fall
but i'm not sure that is the good mean to choose the number of
cluster!

> On Tue, 25 May 2004, n.bouget wrote:
> 
> > Is there a equivalent of the function "eboulis()" (which is a
> > new funtion on Splus) on R? 
> 
> There is no such function in the latest S-PLUS, version 6.2.
 Further,
> S-PLUS is an English product and `eboulis' appears to be French.
> 
> > Otherwise, with which function can we see the best number of
> > cluster we have to choose?
> 
> Please check out where you got that function from, and give
us a 
> description of what you are looking for.  We cannot tell you
the 
> equivalent of a non-existent function (except to give you a
list of 
> non-existent R functions).
> 
> -- 
> 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
> 
> 





More information about the R-help mailing list