Inverse Parametric Analysis of Seismic Permanent Deformation for Earth-Rockfill Dams Using Artificial Neural Networks

Table 3

Comparison of predicted results and actual values of the test samples’ parameters.

	Testing samples	RBF				BP
		𝑐1	𝑐2	𝑐4	𝑐5	𝑐1	𝑐2	𝑐4	𝑐5
1	Actual value	0.0064	0.7500	0.1850	1.3300	0.0064	0.7500	0.1850	1.3300
	Predictive value	0.0064	0.7658	0.1894	1.3263	0.0073	0.8075	0.1851	1.3250
	Error/%	0.0000	2.1067	2.3784	−0.2782	14.5467	7.6702	0.0876	0.3763
2	Actual value	0.0058	0.6750	0.1665	1.1970	0.0058	0.6750	0.1665	1.1970
	Predictive value	0.0062	0.7226	0.1753	1.1950	0.0070	0.7525	0.1675	1.1895
	Error/%	6.8966	7.05185	5.2853	−0.1671	20.7000	11.4833	0.6283	0.6293
3	Actual value	0.0070	0.8250	0.2035	1.4630	0.0070	0.8250	0.2035	1.4630
	Predictive value	0.0068	0.8127	0.1954	1.4354	0.0077	0.8645	0.1982	1.4496
	Error/%	−2.8571	−1.4909	−3.9803	−1.8865	9.7816	4.7920	2.5910	0.9153