Table 2 Evaluation report for offline classification of eight activities (using Neural network model) on the test data set
From: Real-time motion onset recognition for robot-assisted gait rehabilitation
Activity class | Accuracy (%) | Precision (%) | Sensitivity (%) | Specificity (%) | F1-score (%) |
|---|---|---|---|---|---|
Sit to stand | 98.49 | 92.43 | 89.58 | 99.32 | 90.98 |
M.standing | 98.03 | 92.88 | 90.92 | 99.02 | 91.89 |
Stand to sit | 97.73 | 91.48 | 89.39 | 98.86 | 90.42 |
M.sitting | 97.80 | 92.80 | 90.89 | 98.89 | 91.83 |
Walking | 98.19 | 94.70 | 94.58 | 98.92 | 94.64 |
Turning | 97.92 | 93.00 | 89.51 | 99.07 | 91.22 |
Standing | 96.87 | 86.25 | 92.23 | 97.62 | 89.14 |
Sitting | 97.97 | 92.69 | 95.77 | 98.43 | 94.21 |