"[SOLVED]See resulted distance between clusters"

helado

Hello,
I use Rapidminer for performing Clustering (k-means and Hierarchical) on a DataSet using some numerical Similarity Measure. After executing clustering I need to see for all resulted clusters how cluster 1 is similar or dissimilar to cluster 2.
Is there a possibility to get this information?

Thank You!

MartinLiebig

Hi helado,

you got a few different ways. The first is, to have a look at the cluster model. there you can see a centeroid table and have a look where the centeroids are.

The other option is to use the cluster information as label. Then you can do a feature selection (or Weight by Gini Index) to get the most important attributes for your clusters. I would recommend to do a one against all strategy here.

Cheers,

Martin

helado

Thank you very much, Martin!