A couple of weeks ago, I started working through Andrew Ng’s Machine Learning course on Coursera. After working through the Matlab exercises for the first couple of weeks, I decided to try implementing a simple fully connected neural network classifier in Python, using only scipy and numpy. Although I’ll admit that I’m a complete beginner and that the classifier is very simplistic, it was able to achieve 98.76% accuracy in under 120 LOC.
I switched from hosting the source code here to hosting it on GitHub. You can view it here.