Table 3

Parameter setting for TA-NLARNN model and GRNN model

ParameterNLARNN modelGRNN model
Network TypeRecurrent Neural Network (RNN)Recurrent Neural Network (RNN)
Activation FunctionHidden layer activation function:Hyperbolic Tangent (tanh)
Output layer activation function: Linear
Hidden layer activation function:Gaussian
Output layer activation function: Linear
Hidden Units5–2550
Learning Rate0.001–0.010.001–0.1
Training AlgorithmLM. BR, CGF, SCGBR
Input ScalingStandardize or NormalizeNormalize
Loss FunctionRoot Mean Squared Error (RMSE)Root Mean Squared Error (RMSE)
EpochsTypically 100–1,000Typically 50–500
Early StoppingEnabledEnabled
SpreadN/A0–10
Feedback delay1–4N/A
Source(s): Prepared by the authors

or Create an Account

Close Modal
Close Modal