Determine optimal k using Silhouette index

Muhammed_Fatih_

Hi community, 

can someone help me with regard to the question how the k-values of k-Means Clustering approach can be evaluated by using Silhouette index within RapidMiner?

Are there other alternatives to effectively evaluate the k-value beside the x-Means operator in RapidMiner? What would you recommend? 

Thank you in advance! 

Best regards!

lionelderkrikor

Hi @Muhammed_Fatih_

See the answer in your other thread : 

https://community.rapidminer.com/discussion/47497/error-on-determine-silhouette-for-k-means-clustering#latest

Regards,

Lionel

sgenzer

hi @Muhammed_Fatih_ yes please refrain from double-posting. Thank you.

Muhammed_Fatih_

Hi @lionelderkrikor, 

thank you for your hint on the other thread!

Hi @sgenzer, there is no double-posting. This thread targets the calculation of the Silhouette index. Hence, the general question if a possibility exist within RapidMiner to calculate Silhouette? I did not get an answer to my question yet. 

My research showed that there was a plugin from: http://korek.name/web/moje-tvorba/rapidminer-clustering_performance_plugin-average_silhouette-cophenetic_coefficient

But it seems like the link doesn't work anymore. Do you see alternatives for the problem? 

Regards!

[Deleted User]

@Muhammed_Fatih_

Hello

There is a lot of threads about "silhouette"  in community if you search it you can find a lot of information for that
like this one
https://community.rapidminer.com/discussion/17842/how-to-calculate-silhouette-validation-for-k-means-using-r-in-rapidminer

Best regards
mbs