Table 2 ANN hyperparameter values and descriptions
Parameter | Description | Tested values during ANN optimization |
|---|---|---|
Number of inputs | Number of input data variables | 2 |
Number of outputs | Number of output forecasted variables | 1 |
Number of hidden layers | Number of hidden layers | 1 |
Number of hidden neurons | Number of hidden neurons | 2–20 |
Maximum epochs | Max. number of training iterations before training is stopped | 1000 |
Performance goal | The minimum target value of MSE | 0 |
Cross-validation | Cross-validation folds to be used during the grid search | 5 |
Termination Error | The threshold where training terminates when the error is less than or equal to it | 1e-8 |