Table 5: The best settings of network parameters for back-propagation network model.

Network parametersModel FAModel FB

Learning ruleLevenberg-MarquarLevenbergMarquar
Transformation functionsTanh AxonSigmoid Axon
No. of hiding layer22
No. of neurons in hiding layer58
Criterion for terminationMSEmin: 0.0001 MSEmin: 0.0001
Epochsmax: 3000Epochsmax: 3000
Methods for weight update BatchBatch