An implementation of support vector machines in python

* Radial basis kernel - K = e^( -1[(x-y)² / 2sigma²] ) The performance relies on the value chosen for sigma. To chose the optimum sigma that gives the best division of the data set one can implement a cross validation check, a method derived from statistics.

http://en.wikipedia.org/wiki/Support_vector_machine see section Properties http://en.wikipedia.org/wiki/Cross-validation_%28statistics%29

my = 1/2s², my \in (2⁻¹⁵, ... , 2³)