Classification of data as absence of heart disease and presence of heart disease using a simple 2-2-1 Neural Network.
The neural network was developed from scratch in MATLAB using back-propagation with the minimization of the sum of squared error. This was for learning purposes and the classification accuracy in the script is not a proper representation. The main goal of this was to learn how a neural networks work on a fundamental level, including the math behind back-propagation.
Two cases were tested with the network, one was the classification of an XOR problem, and the other was classification of the presence of heart disease.
-- Learning curve for the XOR problem
-- Learning curve for the Heart Diease data