Table 3.

Ten-fold cross-validation accuracies of K-NN and SVM based on HVGE and SE features with optimal parameters

Electrodes	HVGE					SaE
	K-NN		SVM			K-NN		SVM
	k	Acc (%)	C	$\mathit{\sigma }$	Acc (%)	k	Acc (%)	C	$\mathit{\sigma }$	Acc (%)
CP6	24	79.3	8	33.47	79.4	26	75.3	0.3	23.08	76.6
$C$1, $C$3, FC5	9	87.5	8	0.93	87.6	23	84.5	8	0.55	83.8
$13$ channels	9	95.6	8	0.09	95.8	21	89.7	4	0.07	90.2
$63$ channels	7	98.2	4	0.01	98.1	7	95.2	8	0.01	94.3