Research Article
EEG Signal Classification Using Manifold Learning and Matrix-Variate Gaussian Model
Table 4
Cross-validation performance results for different algorithms in Exp.3.
| Feature extraction | Subj.1 (%) | Subj.2 (%) | Subj.3 (%) | Subj.4 (%) | Subj.5 (%) | Subj.6 (%) | Subj.7 (%) | Subj.8 (%) | Subj.9 (%) | Subj.10 (%) | Average (%) |
| None | 43.33 m = 1 | 79.67 m = 1 | 58.31 m = 2 | 50.66 m = 2 | 48.33 m = 2 | 41.00 m = 2 | 57.60 m = 2 | 42.36 m = 1 | 39.33 m = 1 | 84 m = 1 | 54.45 | LDA | 39.33 m = 1 | 73.67 m = 1 | 50.00 m = 1 | 45.00 m = 1 | 43.34 m = 1 | 40.33 m = 1 | 54.77 m = 1 | 39.67 m = 1 | 39.00 m = 1 | 75.33 m = 1 | 50.04 | | | | | | | | | | | 2DLDA | 41.13 m = 1 | 72.33 m = 1 | 52.67 m = 1 | 47.52 m = 1 | 45.67 m = 1 | 39.83 m = 1 | 54.93 m = 1 | 40.33 m = 1 | 39.13 m = 1 | 77.33 m = 1 | 51.09 | | | | | | | | | | | DLPP | 41.67 m = 2 | 75.33 m = 1 | 52.67 m = 2 | 46.00 m = 2 | 49.33 m = 2 | 39.33 m = 1 | 56.00 m = 2 | 39.67 m = 2 | 39.67 m = 1 | 81.67 m = 1 | 52.13 | | | | | | | | | | | 2DDLPP | 41.67 m = 1 | 78.00 m = 1 | 55.33 m = 2 | 47.33 m = 2 | 45.33 m = 1 | 40.67 m = 1 | 55.33 m = 2 | 41.67 m = 2 | 40.33 m = 2 | 79.67 m = 1 | 53.13 | | | | | | | | | | | B2DDLPP | 45.67 m = 1 | 80.67 m = 2 | 59.33 m = 2 | 51.67 m = 2 | 52.82 m = 2 | 45.00 m = 2 | 59.70 m = 2 | 44.40 m = 2 | 42.33 m = 1 | 86.33 m = 1 | 56.79 | | | | | | | | | | |
|
|
For each method and each subject, optimal related to FBCSP’s output and the optimal dimension ( ) are presented. |