k-means Clustering
k (Number of Centroids):
n (Number of Points):
Method to pick starting centroids:
Random Positions
Random Point in Dataset
Distance Measure:
Euclidean Distance (L2-Norm)
L1-Norm
Cosine Distance (with L2)
Method to determine centroids for next iteration:
Mean Point
Method to handle empty clusters:
Random Position
Cluster!
References:
k-means clustering