An Optimized Classification Algorithm by Neural Network Ensemble Based on PLS and OLS

Table 2

Comparison of every classifier’s accuracy for Test 1.

Network model	Weights	Training steps	1996	1997	1998	1999	2000	SSE
Actual value	—	—	1	2	2	2	1	—
Subclassifier I  (BP NN)	0.3179	1036	0.5031	1.8796	2.2593	1.6927	1.1268	0.4392
Subclassifier II  (EBF NN)	0.2311	45	1.8109	2.3207	1.3358	2.2819	0.6651	1.3932
Subclassifier III  (Elman NN)	0.4510	925	1.5218	1.9076	1.8978	2.1038	0.8255	0.3325
PLS-Elman	—	447	1.4129	2.2106	1.9073	1.8775	0.8691	0.2556
Assembling system	—	—	1.2919	1.9942	1.8828	2.0145	0.9451	0.1022