FIRNet Demo 2 -- The Mackey-Glass Equation

FIRNet Demo 2 -- The Mackey-Glass Equation

Input data set has 1500 real vectors
FIR network was trained on first 1300 and used to predict the final 200
Network had 2 layers with 1:5:1 nodes and 7:3 taps
After 400 epochs with a learning rate of 0.012, MSE was 0.000267

Mackey-Glass data set -- first 200 points

Mackey-Glass data set -- final 200 points
Blue is original, red is prediction

The FIRNet Page

Back to Demo #1

Forward to Demo #3