Table 7

Fixed and Adjustable hyperparameters of trained DANNs

Fixed hyperparameterChoices
Input scaler [−]Min-Max
Output scaler [−]Min-Max
Optimization algorithm [−]Adam
Activation function of the hidden layers [−]Rectified Linear Unit
Activation function of the output layer [−]Linear
Loss function [−]MSLE
Dropout [%]25%
Maximum number of epochs [epoch]1,500
Patience for early stopping [epoch]100
Batch size [−]32
Adjustable hyperparameterMinimum valueMaximum value
Number of hidden layers [−]13
Number of nodes per layer [−]8100

or Create an Account

Close Modal
Close Modal