Sharan Basappa <sharan.basappa at gmail.com> writes:
> Can someone please help me to clarify the different between fit and
> predict functions of kmeans?
What is the significance of `predict' in K-means? It is an unsupervised
clustering algorithm. My intuition says that the cluster composition
itself might change if you add a new example and re-run K-means.
On the other hand if you don't want to change the cluster compositions
and just want to find a cluster for a new example then it is not
K-means. In that case it is Knearest classifier.