Table 5.

Performance of different classifiers (average 10-fold CV)

	Model	Accuracy (%)	Precision (%)	Recall (%)	F1 score	AUC	# of features
HOS data	KNN	89.7	90.0	88.5	0.892	0.897	125
	SVM-L	82.3	81.7	81.8	0.817	0.822	200
	SVM-P2	86.7	88.2	84.1	0.861	0.866	200
	SVM-P3	87.8	88.2	86.5	0.872	0.878	225
	SVM-RBF	92.9	92.4	93.0	0.926	0.928	150
	RF	85.4	87.5	81.9	0.843	0.854	175
TS data	KNN	92.6	90.4	94.8	0.925	0.927	2518
	SVM-L	81.0	80.4	81.1	0.806	0.810	975
	SVM-P2	93.1	92.1	94.0	0.930	0.931	875
	SVM-P3	93.2	92.0	94.5	0.932	0.932	950
	SVM-RBF	95.4	96.3	94.3	0.953	0.954	950
	RF	87,8	89.0	85.5	0.872	0.877	800
Raw EEG	Proposed CNN (Oh et al., 2018)	99.2	98.9	99.4	0.992	0.992	-
		95.4	95.2	95.5	0.953	0.954	-