Table 3.

The classification accuracy and F1-score of 10 subjects using five methods in the second stage (%).

Subject	Accuracy					F1-score
	STLDA	DCPM	DeepLDA	EEGNet	STA-CNN	STLDA	DCPM	DeepLDA	EEGNet	STA-CNN
S1	52.7	54.0	63.5	67.9	71.4	28.8	28.1	36.6	30.6	34.2
S2	74.5	72.1	72.0	74.7	79.2	48.9	42.5	41.2	49.2	52.5
S3	54.3	58.6	64.6	60.1	63.0	30.8	32.7	33.0	34.8	34.1
S4	83.1	85.4	87.3	87.7	88.7	64.1	67.7	70.1	71.3	72.8
S5	76.5	74.6	77.8	80.0	83.0	49.2	47.3	48.8	52.2	55.9
S6	61.3	57.0	68.0	65.5	69.6	34.2	34.4	35.7	41.2	39.5
S7	83.2	84.4	84.4	86.6	87.9	65.4	67.2	64.5	68.0	71.2
S8	59.4	58.6	56.3	60.5	65.5	24.5	32.7	28.0	25.7	26.8
S9	71.7	73.1	78.4	78.0	79.9	43.5	46.3	51.1	49.7	52.5
S10	78.3	79.5	87.1	87.7	89.0	60.2	61.9	72.2	71.5	75.5
Mean ± Std	69.5 ± 11.6	69.7 ± 11.8	73.9 ± 10.8	74.9 ± 10.9	77.7 ± 9.8	45.0 ± 15.1	46.1 ± 14.9	48.1 ± 16.0	49.4 ± 16.7	51.5 ± 17.6