Sparse Representation-Based Extreme Learning Machine for Motor Imagery EEG Classification

Table 4

Comparisons of classification results on 2-class and 4-class BCI datasets using different methods.

Datasets		Methods
		ELM	FDDL	FDDL-ELM	H-ELM
Dataset IVa	A1	60.71	57.50	61.70	63.39
	A2	100	84.29	100	98.39
	A3	73.37	70.51	73.88	64.08
	A4	86.61	65.00	88.17	85.67
	A5	79.05	77.14	79.64	85.16
	Mean	79.95	70.89	80.68	79.33
Dataset IIIa	B1	96.89	93.11	97.78	98.56
	B2	68.33	60.33	68.00	60.00
	B3	96.83	96.33	96.83	98.33
	Mean	87.35	83.26	87.54	85.63
Dataset IIa	C1	76.43	62.77	76.74	75.69
	C2	45.38	32.92	45.70	47.74
	C3	76.32	70.76	77.13	76.98
	C4	59.24	45.59	60.50	61.84
	C5	37.12	31.42	36.24	37.85
	C6	45.90	35.28	47.57	47.08
	C7	78.99	66.91	80.30	80.07
	C8	81.87	63.72	80.60	76.46
	C9	67.36	65.48	69.10	76.42
	Mean	63.18	52.76	63.76	64.46