I have a dataset with categorical data, some are binary some are options that have been numerically coded but are not ordinal. Most of this data is categorical "Yes/No" or Nominal for example Favorite Social Media Platform : Facebook, Twitter, Instagram

I want to create clusters that are homogenous within clusters but different between clusters.

What is the best method for this? I am using Hierarchical clustering, but is Between-Groups, Within-Groups, Nearest Neighbour or an alternative method best?

  • $\begingroup$ There is no best method, what have you tried? $\endgroup$ – user2974951 Feb 14 at 13:00
  • $\begingroup$ I have tried all the methods used above as well as K-Modes which i read was useful for data of this nature - but no satisfactory clusters that are heterogenous between are emerging. I suppose the data is just that varied, so it was not a problem with the clustering methodology. $\endgroup$ – Anjali A Feb 28 at 5:20
  • $\begingroup$ This is hard to answer without actually seeing the data. So you may try other clustering methods, such as Gaussian Mixture Models (GMM) or even some classification algorithms such as Support Vector Machines (SVM). $\endgroup$ – user2974951 Feb 28 at 18:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.