Re: [R] Different results in different runs with identical parametersin CLARA

From: TEMPL Matthias <Matthias.Templ_at_statistik.gv.at>
Date: Fri 10 Jun 2005 - 21:05:02 EST

  

> Dear All R Friends,
> When I run my data in any time with the below codes, I
> receive different results.
 

Of course. See in
L. Kaufman and P. Rousseeuw. Finding Groups in Data. John Wiley & Sons, Inc, 1990.
There is a "random part" in clara.  

> My data , k , samples, trace are
> identical in any run.
>
> c<- clara(mydata,4, metric= " euclidean " , stand= TRUE,
> samples=5 , trace=3, keep.data=TRUE , rngR=TRUE)
>
> result of first try:
> Average silhouette width per cluster: 0.5881658
> result of second try:
> Average silhouette width of best sample: 0.6294549
> result of third try:
> Average silhouette width of best sample: 0.6609939
> ...
> I think that only best sample changes in any run.
> The question is here:
> Which try ( or run) is optimal? How many try do I need to
> achive to optimal case? Is it reliable ? Best Regards, Amir
>
 

See it as *Explorative Data Analysis*. Each of your different results give you additional ideas of the structure of your data.  

Best,
Matthias



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 Fri Jun 10 21:19:44 2005

This archive was generated by hypermail 2.1.8 : Fri 03 Mar 2006 - 03:32:29 EST