A guide on how to build a artificial neural net in python with numpy. With detailed step by step explanations for every step.
We generate a set of random binary numbers and then use back propagational neural nets to make an exclusive or (XOR) prediction. You can play around with the number of hidden layers, learning rates and epochs(cycles) to look at the effect that these changes have on the result.
Hopefully this provides a great opportunity to understand some of the basic concepts involved in machine learning.