The demo program is coded using raw Python (no libraries) with the NumPy numeric library, but you should have no trouble refactoring to another language, such as C# or Visual Basic, if you want.

The neural network has (4 * 12) + (12 * 1) = 60 node-to-node weights and (12 + 1) = 13 biases which essentially define the neural network model. Using the rolling window data, the demo program trains ...