Infinite loop

[ASH1998]

when i use another dataset of 4000, 2 shape...the norm>epsilon is always satisfied so it runs endlessly...so if you could use another parameter i.e max_iterations = 300(or some number), the problem would be solved!!
we can use the iterations to be compared with max_iterations and when the limit exceeds we break out of the while loop!!

[jinchuika]

For me, it always falls into an infinite loop (or it seems like) even with the provided dataset and notebook. Am I doing something wrong?

[ASH1998]

naah...just put a break condition after like 300 iterations...(i choose 300 cause its default in sklearn.cluster)
you can use the iterations which is inside the while loop...i think he forgot to do so...the purpose of iterations is that only

[jinchuika]

Good to know. I ran 1000 iterations but didn't know if it was "enough".

[turkialjrees]

k-means can not run infinte and has an upper bound of O^(nkd) in d-dimensional space.

Related defention can be found " in https://www2.cs.duke.edu/courses/spring07/cps296.2/papers/kMeans-socg.pdf