Motor Imagery Recognition Method Based on Multisource Transfer Learning and Multiclassifier Fusion

Table 2

Labeling accuracy of unlabeled samples in multiple source domains based on semi-supervised learning(%).

Subjects	The number of correct labeled samples	The number of mislabeled samples	Accuracy(%)
1	127	10	92.70
2	144	28	83.72
3	184	39	82.51
4	302	34	89.88
5	344	56	86.00
6	508	71	87.74
7	262	25	91.29
8	207	33	86.25
9	254	31	89.12