KNN Algorithm
Given a set of points & a positive integer K > 0 , we return K lists of points (clusters) according to the classical K-Means Algorithm
The initialization step which is done is called KNN ++
The code is improved with two accelerations proposed in the paper :
Phillips S.J. (2002) Acceleration of K-Means and Related Clustering Algorithms. In: Mount D.M., Stein C. (eds) Algorithm Engineering and Experiments. ALENEX 2002. Lecture Notes in Computer Science, vol 2409. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-45643-0_13
Mark run in the target of build.xml & simply execute the ant
Given this set of points :
We obtain :
And after a (very harsh) benchmark of tests :
