X-Means Clustering

Muhammed_Fatih_

Hello community, 

can someone explain how X-means clustering generally works in Rapidminer? I have a problem where the number of clusters k is not known so that I have to determine or rather optimise the number of clusters before going further to the interpretation of the clustering structure.

Is X-means the appropriate Clustering option? Are there alternatives with regard to the optimisation of k?  

I thank you in advance for your help! 

Best regards, 

Fatih 

IngoRM

Hi,

Yes, x-Means finds a good k.  Another approach is to create a graph of a cluster criterion vs. k and looking for the so-called "elbow".  If you search the community for "elbow" you should find some examples.  AutoModel's x-Means also could be helpful for you.

Cheers,
Ingo