perceptron.py implements a perceptron object from scratch using only Numpy as its base module. Features of the Perceptron class include regularization, cross-entropy loss, the delta rule, and modifiable parameters (e.g., the learning rate alpha). Data from Kaggle in the form of medical data on patients with diabetes is included as an example. diabetes.py deploys the perceptron and attains an accuracy of over 72 percent.
While the code is original, the equations were gleaned from Coursera and the following lecture notes.