This is a simple demonstration of the k-means clustering algorithm.

Select some points, and click submit. An example with its solution is already shown.

ERROR: You need a modern browser to use this page (One that supports the "canvas" tag). Try Firefox or Chrome.
Number of classes:

Notes

  • Initialization

    It is initilized by randomly choosing K of the input points and using those as the cluster centers.

  • Termination

    It runs for 20 iterations, or until the labeling does not change. For these demos (small number of points in two dimensions), it often converges in 2-5 iterations.