Attention-Based DSC-ConvLSTM for Multiclass Motor Imagery Classification

Table 4

Within-subject classification accuracy_HGD.

Algorithm	Accuracy (%)														Mean
	A01	A02	A03	A04	A05	A06	A07	A08	A09	A10	A11	A12	A13	A14
FBCSP	76.12	85.92	95.79	94.22	97.73	89.89	88.87	94.90	78.90	89.87	85.87	95.78	95.12	69.70	88.5
Deep ConvNet	75.62	87.50	97.50	98.12	94.37	93.75	86.88	86.25	74.38	94.37	78.56	94.37	98.75	67.50	88.1
Shallow ConvNet	79.37	81.25	96.88	97.50	91.87	92.50	82.50	93.75	76.88	88.75	80.00	93.13	95.63	67.50	87.0
EEGNet-4, 2	82.32	90.63	94.36	95.31	90.34	91.10	90.13	90.99	85.36	92.34	88.33	94.18	95.21	73.10	89.6
EEGNet-8, 2	81.81	90.86	97.63	97.67	93.48	94.46	92.66	91.29	84.66	91.64	87.63	93.48	93.51	70.40	90.1
Ours	85.25	93.55	97.46	98.68	97.47	95.27	93.28	92.44	88.20	93.66	90.65	95.67	97.38	77.79	92.6