Python class that implements genetic machine learning approach to select the fittest vector.
from vectorgenetics import VectorEvolution
def error_function(vector):
return error
ve = VectorEvolution(error_function, vector_shape)
ve.evolve() Also example usage on Erdos problem.
Aim is to find such set of points on plane that there are as much as possible distances of length 1 between them.
Result i got on set of 200 points after 300 generations. ( Computation was intense )
You can find more description at http://allmachineslearn.blogspot.com/2016/12/trying-to-solve-erdos-problem-with.html