Automated Flare Prediction Using Extreme Learning Machine

Table 3

Comparison of all the methods.

Level	Methods	Correctness	True positive	True negative	Weighted true rate	Positive accuracy	Negative accuracy	Weighted accuracy
0	LR	0.83	0.65	0.93	0.85	0.83	0.83	0.83
	SVM	0.85	0.74	0.89	0.85	0.73	0.90	0.85
	LR + SVM	0.86	0.76	0.89	0.86	0.72	0.92	0.86
	LR + ELM	0.81	0.64	0.87	0.81	0.65	0.86	0.81
1	LR	0.76	0.71	0.77	0.75	0.26	0.96	0.76
	SVM	0.70	0.43	0.74	0.65	0.18	0.9	0.70
	LR + SVM	0.72	0.53	0.74	0.67	0.14	0.95	0.72
	LR + ELM	0.70	0.44	0.75	0.67	0.29	0.85	0.70
2	LR	0.74	0.83	0.73	0.76	0.15	0.93	0.74
	SVM	0.68	0.39	0.72	0.62	0.13	0.93	0.68
	LR + SVM	0.65	0.35	0.72	0.61	0.22	0.83	0.65
	LR + ELM	0.70	0.48	0.74	0.66	0.28	0.87	0.70
3	LR	0.86	1	0.86	0.88	0.03	1	0.86
	SVM	0.83	0.37	0.87	0.80	0.21	0.94	0.83
	LR + SVM	0.84	0.44	0.87	0.81	0.21	0.95	0.84
	LR + ELM	0.85	0.52	0.89	0.84	0.35	0.94	0.86