I've been using Orange for a while to rapidly prototype a few classification models. One of the ones I've been using is 'Naive Bayes'. If I understand correctly, there are a few types available based on the Bayesian principle -> Gaussian, Multinomial, Bernoulli etc. I see these as explicitly classes with scikit-learn as well.

However I'm unable to tell which type is the one 'Orange' adopts for its 'Naive Bayes' widget. Can someone help me with this?

  • $\begingroup$ Welcome to DataScienceSE. I don't know Orange but since the " NB type" depends completely on the type of data for the features, my guess would be that Orange automatically selects the appropriate version. For example if the features are continuous then it selects Gaussian, if they are boolean then Bernouilli... $\endgroup$
    – Erwan
    Jul 9, 2021 at 12:19
  • $\begingroup$ Thanks Erwan, appreciate it! That makes sense, I will experiment with the various forms and see if my results match with Orange's accordingly. $\endgroup$ Jul 9, 2021 at 18:18


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.