Hacker News new | past | comments | ask | show | jobs | submit login

Sorry, I don't understand your argument. Could you clarify what you mean by "everything"? Is there another clustering algorithm you're thinking of?



i just mean that k-means is a rather simple algorithm for finding centroids. most algorithms that are more sophisticated might do better than k-means.


Sorry for the late response. And actually, despite the plethora of clustering algorithms, k-means is still a very commonly-used out-of-the-box technique! (Probably for its simplicity and existing popularity)




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: