EURASIP Journal on Advances in Signal Processing

Table 3 Performance comparisons for the IS and VC problems

From: An incremental learning algorithm for the hybrid RBF-BP network classifier

Data sets	Method	N _H neurons	Training time(s)	Testing η _o	Testing η _a
IS	SVM	96^a	11.61	90.62	–
	MRAN	78	11.68	85.82	–
	GAP-RBF	87	5.77	86.34.	–
	ELM	49	0	90.23	–
	OS-ELM	100	0.01	90.67	–
	ILRBF-BP	77 and 8^b	2.09	91.57	–
VC	SVM	234^a	10.74	68.72	67.99
	MRAN	105	10.38	60.24	60.02
	GAP-RBF	81	9.87	58.94	58.17
	ELM	300	0.09	68.01	67.39
	OS-ELM	300	0.12	68.95	67.56
	ILRBF-BP	258 and 9^b	11.53	70.17	69.43

^aSupport vectors
^bRBF and BP hidden neurons

Back to article page