K-means implementation in Python using NumPy library

By Prateek Kumar

Implementation of most famous clustering algorithm(K-means) using Expectation(E)-Maximization(M) technique in Python using NumPy.

The given packet contains kmeans_algo module. KMeans class contains implementation of K-means algorithm in OOP way.

methods-

dist(v1,v2)- It calculates and return Euclidean distance between two vectors(v1 and v2).

train(X,k,epochs)-

parameters:

X: a (m,n) shaped NumPy array consisting of m training examples with n attributes.

k: numbers of cluster centres

epochs: Number of times E-step and M-step need to be repeated.

 

return:

cluster:A Python dictiionary with keys from 0 to k-1.
cluster[i] is a python dictiionary where i denotes ith cluster consisting of two parameters -
'centre': denoting the mean of that all points belonging to that cluster
'points': denoting a list of all the points that falls in that cluster