Difference between K-means and Kernel K-means

K-means:

K-Means aims to partition N observations into K clusters in which each observation belongs to the cluster with the nearest mean (cluster centroid).

Kernel K-means:

It can be applied when the objective function can be written as a function of dot products. It allows us to map our current feature vectors into higher dimensional spaces in a more computationally efficient manner using the kernel trick.

K-means vs Kernel K-means:

K-means	Kernel K-means
1. K-means clustering on the spectral embedding of affinity matrix.	1. kernel K-means clustering directly to the affinity matrix.
2. CPU and GPU aren't parallelized K-means	2. CPU and GPU parallelized kernel K-means
3. K-means dataset performs good	3. Kernel K-means dataset performs poorly