An Empirical Comparison of Multiple Linear Regression and Artificial Neural Network for Concrete Dam Deformation Modelling

Table 2

Statistical performance of the MLR, SR, BP, and ELM single-point models.

DAM	model	MAE		MSE		S		R		Time; (s)
		Training	Predicting	Training	Predicting	Training	Predicting	Training	Predicting
DJ	MLR	0.5194	0.4962	0.4710	0.5599	3.3345	3.2524	0.9519	0.9054	1.3800
	SR	0.5210	0.6517	0.4999	0.6951	3.4872	3.1291	0.9489	0.9143	1.3910
	BP	0.4325	0.5889	0.3595	0.7087	3.7087	3.4262	0.9649	0.9360	2.4140
	ELM	0.4240	0.4555	0.3504	0.5005	3.3246	2.4620	0.9652	0.9420	1.4360
FM [17]	MLR	0.8355	1.0673	0.9947	1.7170	2.0376	2.8455	0.9388	0.9190	2.1619
	SR	0.9097	1.0959	1.1750	1.6609	2.5263	2.3165	0.9361	0.9099	2.7656
	BP	0.4270	0.8491	0.3505	1.1311	2.2703	2.3678	0.9824	0.9466	5.3310
	ELM	0.3978	0.7020	0.2653	0.8929	1.6336	2.2076	0.9840	0.9564	2.9265

MAE = mean absolute error; MSE = mean square error; S = maximum absolute error; R = the correlation coefficient; MLR = multiple linear regression; SR = stepwise regression; BP = backpropagation; ELM = extreme learning machine. DJ = the Dongjiang arch dam; FM = the Fengman gravity dam.