Can k-means fall into an infinite loop ?
-
29-09-2019 - |
Question
I've studied the k-means algorithm and I know how it works.
Just curious,is there any situation that this algorithm will go into an infinite loop,say if we have some particular bad choices for initial centroid points? I could only imagine a situation k-means will get to local minimum with bad initial choices.
Solution
No. k-means has an upper bound of O(nkd) in d-dimensional space.
Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow