Application of BP Neural Network Improved by Fireworks Algorithm on Suspender Damage Prediction of Long-Span Half-Through Arch Bridge

Table 9

Prediction results of each model for comparative analysis.

Test no.	Model	AE (%)	AEmax (%)	R	T/s
1	BPNN	5.74	20.81	0.902	60.49
	GA-BPNN	4.26	13.26	0.948	52.36
	PSO-BPNN	4.03	13.19	0.959	51.42
	FWA-BPNN	3.59	9.79	0.963	51.73
2	BPNN	5.62	22.35	0.913	59.16
	GA-BPNN	4.11	12.79	0.962	43.69
	PSO-BPNN	3.98	11.36	0.969	42.58
	FWA-BPNN	3.24	9.14	0.975	40.48
3	BPNN	5.49	20.10	0.926	51.38
	GA-BPNN	3.83	11.51	0.964	41.67
	PSO-BPNN	3.72	11.34	0.973	41.09
	FWA-BPNN	3.07	8.82	0.982	40.16