0

I have a dataset with coordinates and my goal is to find the optimal K so I can cluster using K-Means.

I've already tried using Kneedle method via https://github.com/lukehb/137-stopmove/blob/master/src/main/java/onethreeseven/stopmove/algorithm/Kneedle.java but I can't seem to get it to work due to dependency issues.

What else can I try?

Vahe Karapetyan
  • 149
  • 2
  • 9
  • 2
    @Augusto It's not really software engineering. It's rather data science. And ... tadaaa... here it is: https://datascience.stackexchange.com/questions/5656/k-means-what-are-some-good-ways-to-choose-an-efficient-set-of-initial-centroids - similar questions are found on other network sites, e.g. https://stats.stackexchange.com/questions/277007/rule-of-thumb-on-the-best-k-in-k-means-clustering - but also here on stackoverflow: https://stackoverflow.com/questions/1793532/how-do-i-determine-k-when-using-k-means-clustering (Anyone opposed to closing this one as a duplicate of the latter?) – Marco13 Jun 18 '19 at 14:53
  • 1
    How about using VRC criterion? Or silhouette? Or Dunn? – Has QUIT--Anony-Mousse Jun 18 '19 at 19:32

0 Answers0