From: Hemodynamic functional connectivity optimization of frequency EEG microstates enables attention LSTM framework to classify distinct temporal cortical communications of different cognitive tasks
Hyperparameters
Tuned parameters
Hidden layer size
256
Batch size
64
Training epoch numbers
1000
Rate dropout
Input Layer: 0, 1st LSTM Layer: 0.2, 2nd LSTM Layer: 0.1, 3rd LSTM layer: 0.2
Recurrent depth
3
Learning rate
0.001