Clustering can be done using the following methods:
1) Partitioning methods, e.g. k-means.
2) hierarchial methods
3) Model based.
k -means method:
1) Select k random elements which act as the centroid of the initial clusters.
2) Each point in the given data calculates the distance from these points and according to the distance new clusters are formed.
3) The centroid of the resulting clusters are calculated again. These centroids are the new "k"
4) Do step 2 and 3 again. Till the clusters converge. i.e. the centroids of the new clusters do not change after iterations.
