Table 3.
Evaluation matrix for one-channel analysis for EMG data using different feature selection techniques.
| Muscle | Feature Ranking Technique | Features | Accuracy (%) | Precision (%) | Sensitivity (%) | F1 Score (%) | AUC |
|---|---|---|---|---|---|---|---|
| GL | Chi-Square | Top 8 | 81.30 | 82.22 | 81.30 | 81.54 | 0.83 |
| mrmr | Top 11 | 80.92 | 81.59 | 80.92 | 81.11 | 0.82 | |
| Relieff | Top 7 | 84.73 | 84.89 | 84.73 | 84.76 | 0.85 | |
| fscnca | Top 10 | 81.30 | 82.27 | 81.30 | 81.54 | 0.83 | |
| TA | Chi-Square | Top 8 | 80.53 | 80.91 | 80.53 | 80.59 | 0.82 |
| mrmr | Top 9 | 84.35 | 85.06 | 84.35 | 84.48 | 0.87 | |
| Relieff | Top 9 | 81.30 | 81.58 | 81.30 | 81.40 | 0.82 | |
| fscnca | Top 11 | 83.59 | 83.93 | 83.59 | 83.67 | 0.86 | |
| VL | Chi-Square | Top 10 | 74.81 | 75.55 | 74.81 | 75.09 | 0.80 |
| mrmr | Top 11 | 75.95 | 76.77 | 75.95 | 76.24 | 0.82 | |
| Relieff | Top 8 | 77.10 | 77.96 | 77.10 | 77.39 | 0.83 | |
| fscnca | Top 11 | 76.72 | 77.25 | 76.72 | 76.84 | 0.83 |