[R] K-means result - variance between cluster

[Christian Hennig]

Dear Ralph,

between and within clusters sum of squares (if you want variances, you 
need to divide them by the appropriate constant!) add up to the 
overall sum of squares, so you can get the beween clusters ss by 
computing the overall ss (one possibility to get this is to run kmeans 
with k=1) and subtracting the within cluster ss from it.

Note, however, that the F-value cannot be interpreted in the usual way and 
is particulary not F-distributed when computed on clusters from k-means, 
because for F-distribution you'd need to assume that groups are determined 
independently of the data.

Hope this helps,
Christian

On Fri, 2 Jul 2010, Ralph Modjesch wrote:

> Hi,
>
> I like to present the results from the clustering method k-means in
> terms of variances: within and between Cluster. The k-means object
> gives only the within cluster sum of squares by cluster, so the between
> variance part is missing,for calculation the following table, which I
> try to get.
>
> Number of | Variance within | Var between | Var total | F-value
> Cluster k | cluster         | cluster     |           |
> ===============================================================
> 2 .......| 25,00 ..........| 75,00 ......| 100 ......| 1,5
> 3 .......| 45,00 ..........| 55,00 ......| 100 ......| 1,7
>
> Is there any package/ function which will do that?
>
>
> --
> Mit freundlichen Grüßen
>
> Ralph Modjesch
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>

*** --- ***
Christian Hennig
University College London, Department of Statistical Science
Gower St., London WC1E 6BT, phone +44 207 679 1698
chrish at stats.ucl.ac.uk, www.homepages.ucl.ac.uk/~ucakche