When it's better to use K-Medoids rather than K-Means? Can anybody give some examples of dataset for the same?

  • $\begingroup$ From Wikipedia: It is more robust to noise and outliers as compared to k-means because it minimizes a sum of pairwise dissimilarities instead of a sum of squared Euclidean distances. $\endgroup$ Jul 9 '20 at 7:40
  • 2
    $\begingroup$ You may also want to use K-medoid when you want the representers of the classes to be one of the points in your dataset. Ex : representer of a class of pictures of faces is a blurry picture for K-means and it is a real picture of someone's face in K-medoid, one of the exemples of the dataset. $\endgroup$
    – TMat
    Jul 9 '20 at 8:00
  • 1
    $\begingroup$ The question is similar to when to use median instead of mean. Check the answers here stats.stackexchange.com/questions/6913/… Mean is trying to minimize L2 loss and Median is trying to minimize L1 loss. stats.stackexchange.com/questions/34613/… $\endgroup$
    – Haitao Du
    Jul 9 '20 at 8:04

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.