Research Article
Investigating Feature Ranking Methods for Sub-Band and Relative Power Features in Motor Imagery Task Classification
Table 1
Classification accuracy of different classifiers.
| Subjects | Classification accuracy (in %) | Tree | Fine KNN | Weighted KNN | Quadratic SVM | Random forest |
| S01 | 89.1 | 92.9 | 94.3 | 65.3 | 89.4 | S02 | 87.3 | 91.9 | 92.8 | 70.5 | 91.1 | S03 | 89.5 | 92.9 | 94.2 | 71 | 91.34 | S04 | 88.2 | 90.6 | 93.6 | 67.3 | 90.625 | S05 | 85.5 | 91.8 | 93.1 | 66.1 | 86.77 | S06 | 87.7 | 90.2 | 93.1 | 58.4 | 90.86 | S07 | 88.7 | 89.1 | 91.6 | 69 | 87.01 | S08 | 86.7 | 91.3 | 92.9 | 64.4 | 89.66 | S09 | 92 | 92.7 | 95.2 | 73.7 | 92.78 | S10 | 87.7 | 92.9 | 94.3 | 67.5 | 91.34 | AVG | 88.24 | 91.63 | 93.51 | 67.32 | 90.0885 |
|
|
Bold letters show the maximum classification accuracy of the classifier.
|