Prediction of Compressive Strength Behavior of Ground Bottom Ash Concrete by an Artificial Neural Network

Table 4

The adjustment parameters used in the ANN models.

Adjustment parameters	ANN models
Number of input layer units	4
Number of first hidden layer units	4 : 1 : 14
Number of second hidden layer units	1 : 1 : 2
Number of output layer units	1
Number of epochs	500 : 500 : 20,000
Performance goal	1.00E − 06
Momentum rate	Default
Learning rate	Default
Training algorithm	BR
Activation function in hidden layer	Log-sigmoid and tan-sigmoid
Activation function in output layer	Purelin linear
Network’s performance	sse