Vector quantisation demonstration

A simple demonstration of vector quantisation

Starting with 4 random points this program demonstrates vector quantisation.
1. First 4 random points are chosen as cluster means.
2. Then we find the cluster mean that is closest to a given point and allocate it to that corresponding cluster. Here, we can set the limits on the trust region (based on Euclidean distance). Say, points which are less than euclidean distance of 2 from a cluster mean belong to this particular cluster.
3. Once the allocation is complete we go on to calculate the next set of cluster means.
4. This is again followed by another round of allocation. We can see how the the clusters and the cluster means change over iterations.
5. I have set the number of iterations at 15. This is just to show how the allocation freezes after a while, once the optimum cluster centres are identified.