How to determine the number of clusters with "SOM" for K-means

mahsasalehi000

hello, I use a data set with 4 coloumn and 955 row. and I need to determine the best number of cluster for K-means by using "SOM" (Self-Organizing Map).

I'm new in rapidminer. please describe how use operators in Process with a picture.ُ

Solving this problem is very necessory and don't have enough time.

please help me. thank you.

Telcontar120

If you search in the community forum, there are a number of other threads that relate to optimizing the value of k for k-means clustering, such as this one: http://community.rapidminer.com/t5/RapidMiner-Studio-Forum/Dynamically-determine-number-of-clusters-k-means/m-p/39731#M27121

Why does the selection of k need to be done via a SOM?  There is a SOM extension available in the marketplace (just search for it) so if you want to go down that road, it is a theoretical option.  But X-means is probably a much faster and more direct approach to the problem.