Backpropagation of error is a learning algorithm for artificial neural networks (typically feedforward) popularized by Geoffery Hinton (PhD, University of Toronto). Although it is not biologically plausible, it serves as a powerful algorithm for solving complex problems with artificial neural networks.
Below is an implementation of a backpropagation softmax network in python. The script would run faster if numpy arrays are used. More detailed descriptions will be added later.
