How do you test your results for overfitting in a k-means run? Some people have said use a training set. I have about 1500 records and about 20 fields.

  • 1
    $\begingroup$ You can't overfit K-Means. It can, however, be non-robust. It's unsupervised learning, not supervised. Keywords for theory: silhouette analysis, elbow method, gap statistics, mutual information $\endgroup$ Mar 26 '19 at 20:03
  • $\begingroup$ @CarlRynegardh although this answer shares the same opinion, this answer seems quite reasonable too. We can settle for a subjective overfitting I think! $\endgroup$
    – Esmailian
    Mar 26 '19 at 20:43

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.