ANN hyperparameter setup used in NARX
| Items . | Value . | Remarks . |
|---|---|---|
| Number of neurons | [5, 5] | Using two hidden layers |
| Learning rate | 0.01 | Using a constant learning rate |
| Max epoch | 1,000 | Interaction using early stopping |
| Solver | ‘Adam’ | Reference (Kingma & Ba 2014) |
| Activation function | Relu | Rectified linear unit |
| Items . | Value . | Remarks . |
|---|---|---|
| Number of neurons | [5, 5] | Using two hidden layers |
| Learning rate | 0.01 | Using a constant learning rate |
| Max epoch | 1,000 | Interaction using early stopping |
| Solver | ‘Adam’ | Reference (Kingma & Ba 2014) |
| Activation function | Relu | Rectified linear unit |