Table 2.
Average recognition results accuracies, sensitivities and specificities (standard deviation in parentheses) using regression and classification model and sensor combinations.
| Regression | |||
|---|---|---|---|
| Sensors | Balanced accuracy | Sensitivity | Specificity |
| ACC+EDA+ST+BVP | 89.0 (13.3) | 86.9 (22.7) | 93.6 (5.8) |
| EDA+ST+BVP | 75.0 (16.0) | 71.6 (27.3) | 90.9 (9.2) |
| EDA+BVP | 75.1 (15.8) | 73.5 (24.5) | 88.3 (7.8) |
| BVP+ST | 82.3 (17.0) | 76.6 (23.8) | 91.0 (6.1) |
| EDA+ST | 72.3 (16.7) | 70.9 (24.2) | 84.0 (11.0) |
| EDA | 73.5 (15.2) | 72.0 (23.7) | 87.5 (10.8) |
| BVP | 78.5 (16.7) | 72.1 (23.7) | 89.0 (7.2) |
| ST | 68.6 (11.8) | 56.1 (19.3) | 83.5 (9.0) |
| ACC | 94.3 (5.5) | 93.2 (7.0) | 96.2 (3.4) |
| Classification | |||
| Sensors | Balanced accuracy | Sensitivity | Specificity |
| ACC+EDA+ST+BVP | 85.2 (14.2) | 82.4 (22.9) | 91.7 (9.4) |
| EDA+ST+BVP | 65.4 (22.8) | 61.6 (33.9) | 76.7 (15.7) |
| EDA+BVP | 69.3 (16.9) | 57.9 (33.3) | 78.7 (11.4) |
| BVP+ST | 74.1 (16.7) | 66.0 (33.5) | 85.6 (10.9) |
| EDA+ST | 64.6 (23.1) | 54.1 (31.8) | 73.2 (17.5) |
| EDA | 64.5 (14.6) | 55.8 (23.9) | 78.1 (12.8) |
| BVP | 70.4 (20.4) | 64.0 (38.2) | 82.9 (11.0) |
| ST | 61.3 (14.9) | 50.5 (27.6) | 76.5 (12.0) |
| ACC | 90.2 (10.1) | 91.4 (9.2) | 96.0 (6.4) |