Implementation of the backpropagation algorithm in Python using only the numpy library.
The effectiveness of the implementation is verified by proving the universal approximation theorem: any neural network architecture can find any mathematical function y= f(x) that can map attributes x to output y
