MCGNet+: an improved motor imagery classification based on cosine similarity

Brain Informatics

Table 2 The performance comparison of the state-of-the-art approaches on the SMR dataset

Model	Accuracy	F1-score	Precision
SVM	0.3488	0.3485	0.3486
Deep ConvNet	0.3507	0.3191	0.4148
FBCSP	0.3511	0.3366	0.3714
RF	0.4008	0.3996	0.4004
EEGNet	0.4616	0.4838	0.5095
Shallow ConvNet	0.4857	0.4789	0.4978
MutualGraphNet (ours)	0.5190	0.5175	0.5208
MCGNet\(^{+}\) (ours)	0.5227	0.5239	0.5278