Precision of Random Forest algorithms for each activity class/intensity for two cases in three phone locations; right hand, right pocket and backpack
| Case 3 | Backpack precision | Pocket precision | Hand precision |
| 7MET | 0.98 | 0.98 | 0.98 |
| 5MET | 0.83 | 0.83 | 0.83 |
| 3MET | 0.95 | 0.95 | 0.93 |
| Walking | 0.94 | 0.96 | 0.95 |
| Sitting | 0.96 | 0.97 | 0.96 |
| Lying | 0.97 | 0.98 | 0.98 |
| Case 4 | Backpack precision | Pocket precision | Hand precision |
| 7MET | 0.87 | 0.85 | 0.85 |
| 5MET | 0.84 | 0.84 | 0.82 |
| 3MET | 0.65 | 0.64 | 0.60 |
| Walking | 0.63 | 0.64 | 0.62 |
| Sitting | 0.62 | 0.62 | 0.65 |
| Lying | 0.70 | 0.70 | 0.75 |
Case 3 includes raw accelerometer data and 58 features. Case 4 includes the vector magnitude of activity counts and 58 features.
MET, Metabolic Equivalent Task.