Research Article
Motor Imagery EEG Decoding Based on New Spatial-Frequency Feature and Hybrid Feature Selection Method
Table 1
The classification accuracy achieved by various feature extraction methods (data set 1).
| Subject | CSP | CSP-FB | SFBCSP | SBLFB | CSP-LBP | CSP-FBLBP |
| L versus R | 77.62 | 79.01 | 76.39 | 77.01 | 77.47 | 77.78 | L versus F | 82.41 | 85.88 | 79.32 | 80.25 | 82.33 | 87.58 | L versus T | 85.26 | 85.42 | 82.87 | 81.25 | 85.80 | 85.65 | R versus F | 81.02 | 84.72 | 76.93 | 76.62 | 84.10 | 86.96 | R versus T | 82.95 | 81.79 | 79.78 | 79.32 | 83.34 | 85.80 | F versus T | 72.61 | 72.61 | 71.30 | 70.68 | 74.46 | 77.55 | Mean ± Std | 80.31 ± 14.25 | 81.57 ± 12.56 | 77.76 ± 12.06 | 77.52 ± 12.03 | 81.25 ± 12.73 | 83.55 ± 12.32 |
|
|