For supervised learning, we know the correct answers for samples, model selection is more easier, we can use k-ford cross validation (this site!) and etc.

But for unsupervised learning, e.g. clustering, how can we determine the optimal number of clusters?

I read some literature about Kullback-Leibler information criterion but its performance is not quite good after I tried some simulations.

Thanks, guys!


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.