Table 1.

Application of three optimization algorithms to fit the coefficients for the ANN model

Optimization algorithm	Function	Epoch	Architecture	RMSE	R2	MAPE (%)	Computing time (s)
Levenberg–Marquardt	trainlm	1000	1-3-1	0.2290	0.9999	0.39	46.23
BFGS quasi-Newton	trainbfg	1000	1-3-1	0.2165	0.9999	0.33	118.72
Batch gradient descent	traingd	1000	1-6-1	0.7722	0.9992	1.38	268.75