Research Article
Comparison of Different Features and Classifiers for Driver Fatigue Detection Based on a Single EEG Channel
Table 1
Optimal combination for different subjects.
| Subject | Optimal combination | Highest Acc | AUC |
| 1 | FE + KNN | 94.3% | 0.976 | 2 | FE + KNN | 86.4% | 0.929 | 3 | FE + RF | 93.4% | 0.981 | 4 | FE + RF | 91.0% | 0.969 | 5 | FE + RF | 92.6% | 0.976 | 6 | FE + RF | 91.3% | 0.974 | 7 | FE + RF | 91.4% | 0.968 | 8 | FE + RF | 92.7% | 0.981 | 9 | FE + RF | 94.4% | 0.983 | 10 | FE + RF | 91.9% | 0.975 | 11 | FE + RF | 90.5% | 0.967 | 12 | FE + RF | 93.2% | 0.979 |
|
|