0

I have data where all features are nominal. I applied SMOTE-NC, then I found that it only works with a combination of nominal and continuous features!.

There is a technique called SMOTE-N (to deal with only nominal features) in the same paper of SMOTE technique but I can't find any code or function for it in python, is there any application or something similar?. or is there any other over or under-sampling technique that works with only categorical features

desertnaut
  • 57,590
  • 26
  • 140
  • 166
Hanan
  • 33
  • 1
  • 5
  • If you have only nominal features, random oversampling is a far easier approach. – DejaVuSansMono Jul 06 '20 at 16:45
  • 1
    The duplicate answers your first question; regarding your second one, please notice that recommendation requests for off-site resources are explicitly off-topic - see [What topics can I ask about here?](https://stackoverflow.com/help/on-topic) – desertnaut Jul 06 '20 at 16:45
  • @DejaVuSansMono I can't find any reference saying it doesn't contradict with having all categorical features. can you please share if you have such information. thanks in advance – Hanan Jul 06 '20 at 16:59
  • @desertnaut I edited the question based on that. now I'm looking for any other technique that could work with only categorical variables. – Hanan Jul 06 '20 at 17:01
  • This question is now closed, plus, despite your editing, you are still asking for a resource (only now you are additionally asking for any other technique). Please open a new question, making sure that it is indeed on-topic (I'm rolling back the current one to its previous form). – desertnaut Jul 06 '20 at 17:14
  • @desertnaut I need to wait for two days to ask another question. I deleted the part which you mentioned it is duplicate. this is quite urgent for me! please let know the solution to post another question or allow this one after the edits – Hanan Jul 06 '20 at 17:18

0 Answers0