Model Construction
•
Neural Network (4-8-8-1) using input
mask
111100
•
Trained using the BFGS algorithm on
800 samples to the MSE predicted by
the Gamma Test (0.00032)
•
MSE on 100 unseen samples 0.00040