On the Training Algorithms for Artificial Neural Network in Predicting the Shear Strength of Deep Beams

Table 4

Summary of different quality assessment criteria over 300 simulations with different ANN structures.

Criteria	ANN-QN		ANN-CG		ANN-GD
	Train	Test	Train	Test	Train	Test
R
Min	0.976	0.705	0.984	0.854	0.954	0.764
Average	0.989	0.961	0.994	0.971	0.993	0.966
Max	0.995	0.993	0.997	0.993	0.997	0.994
Std	0.003	0.032	0.002	0.020	0.003	0.032
RMSE
Min	12.28	14.51	9.37	14.02	9.20	14.15
Average	17.49	32.80	12.86	28.07	13.76	29.44
Max	24.73	71.99	18.06	76.57	28.84	66.07
Std	2.23	8.62	1.51	8.44	2.04	9.01
MAE
Min	9.18	11.07	7.28	10.60	7.21	11.25
Average	13.01	22.56	9.68	19.26	10.23	20.28
Max	17.52	41.24	13.84	38.72	18.80	41.54
Std	1.61	4.89	1.17	4.75	1.40	5.18
MAPE
Min	5.03	6.00	3.72	5.79	4.02	5.97
Average	7.02	12.78	5.29	11.05	5.52	11.35
Max	9.61	28.26	7.70	29.70	10.33	26.58
Std	0.92	3.81	0.69	3.45	0.79	3.50