A simple implementation of the standard k-means clustering algorithm: http://en.wikipedia.org/wiki/K-means_clustering. K-means clustering partitions points into clusters, with each point belonging to the cluster with th nearest mean. As the general problem is NP hard, the standard algorithm, which is relatively rapid, is heuristic and not guaranteed to converge to a global optimum. Varying the input order, from which the initial clusters are generated, can yield different results. For degenerate and malicious cases, the algorithm may take exponential time.