Research Article
Simultaneous Channel and Feature Selection of Fused EEG Features Based on Sparse Group Lasso
Table 5
Experimental results on different datasets.
| Channel/feature selection method | Dataset A | Dataset E | Dataset F | Dataset G | Channel | Feature | Error rate (%) | Channel | Feature | Error rate (%) | Channel | Feature | Error rate (%) | Channel | Feature | Error rate (%) |
| Elastic Net | 44 | 120 | 32.60 | 41 | 89 | 16.16 | 27 | 38 | 43.86 | 49 | 168 | 23.50 | Lasso | 27 | 43 | 31.28 | 38 | 60 | 15.28 | 27 | 38 | 43.86 | 22 | 49 | 23.51 | Sparse Group Lasso (α = 0.25) | 16 | 890 | 31.72 | 22 | 1193 | 17.03 | 13 | 742 | 40.96 | 12 | 868 | 22.22 | Sparse Group Lasso (α = 0.5) | 18 | 737 | 29.26 | 18 | 703 | 16.59 | 14 | 584 | 40.52 | 14 | 686 | 20.94 | Sparse Group Lasso (α = 0.75) | 17 | 440 | 32.15 | 23 | 508 | 15.28 | 14 | 329 | 40.52 | 14 | 370 | 20.94 | Group Lasso | 17 | 1190 | 31.28 | 26 | 1820 | 17.47 | 16 | 1120 | 40.35 | 10 | 700 | 22.22 |
|
|