[R] Re: Validation of Clustering

[Don Eduardo Miranda]

For Clustering validation it maybe good to refer to "On Clustering
Validation Techniques" by Maria Hakidi.  It presents a good survey on these
kind of techniques.  You can even find this an orticles on the subject here
http://www.db-net.aueb.gr/mhalk/Publ_maria.htm
______________
Eduardo Miranda
Departamento de Informática da FCT/ UNL
Quinta da Torre, 2829-516 Caparica, Portugal
Tel: +351-21 294 85 36  - Ext. 10731
Fax: +351-21 294 85 41
E-mail: miranda at di.fct.unl.pt
http://ctp.di.fct.unl.pt/QUASAR

> Date: Fri, 31 Jan 2003 12:34:47 +0100 (CET)
> From: Christian Hennig <hennig at stat.math.ethz.ch>
> To: Francisco do Nascimento Junior <fnj at cin.ufpe.br>
> cc: R-help <R-help at stat.math.ethz.ch>
> Subject: Re: [R] Validation of clustering
>
> Hi,
>
> I am not really sure what you want to know.
> If you simply want to know how to find the resulting clustering, consider
> help(clara.object). For validation and assessment of the clustering, you
> might use a silhouette plot, see help(plot.partition),
> help(partition.object).
>
> For more descriptions consider the book of Kaufman and Rousseeuw; some
free
> material is on
> http://win-www.uia.ac.be/u/statis/index.html
> (Programs->Clustering)
>
> Best,
> Christian
>
> On Thu, 30 Jan 2003, Francisco do Nascimento Junior wrote:
>
> > Hi,
> > I'm using the library cluster to cluster a set of figures (method
CLARA).
> > Somebody that it work with clustering would know informs what I make to
> > evaluate the clustering?
> >
> > Tks VM,
> > Francisco.
> > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> > Francisco Júnior,
> > Computer Science - UFPE-Brazil
> > "One life has more value that the
> > world whole"
> > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^