Table 5 \(\Delta\)TO and \(\Delta\)HS of alternative algorithms for each cohort
| Â | Â | CS | MS | EVF | |||
|---|---|---|---|---|---|---|---|
| Â | Â | Value | p-value | Value | p-value | Value | p-value |
Test 1 | F1-score (%) | 100 | – | 99.1 | – | 95.7 | – |
|  | Recall (%) | 100 | – | 99.4 | – | 96.1 | – |
|  | Precision (%) | 100 | – | 98.8 | – | 95.4 | – |
|  | Mean \(\Delta\)TO [IQR] | 8 [3–15] | 0.801 | 30 [13-77] | < 0.0001 | 31 [17–70] | < 0.0001 |
|  | Mean \(\Delta\)HS [IQR] | 8 [3–18] | < 0.0001 | 19 [8-44] | < 0.0001 | 38 [15–80] | < 0.0001 |
Test 2 | F1-score (%) | 56.4 | – | 49.2 | – | 57.2 | - |
|  | Recall (%) | 56.9 | – | 48.6 | – | 57.1 | – |
|  | Precision (%) | 55.9 | – | 49.6 | – | 57.4 | – |
|  | Mean \(\Delta\)TO [IQR] | 17 [7-71] | < 0.0001 | 50 [18-177] | < 0.0001 | 58 [23–177] | < 0.0001 |
|  | Mean \(\Delta\)HS [IQR] | 13 [5-67] | < 0.0001 | 34 [10-166] | < 0.0001 | 60 [20–171] | < 0.0001 |
Test 3 | F1-score (%) | 97.9 | – | 98.2 | – | 95.2 | – |
|  | Recall (%) | 97.9 | – | 98.6 | – | 95.5 | – |
|  | Precision (%) | 97.9 | – | 98.0 | – | 94.9 | – |
|  | Mean \(\Delta\)TO [IQR] | 33 [15–90] | < 0.0001 | 42 [18–90] | < 0.0001 | 57 [27–100] | < 0.0001 |
|  | Mean \(\Delta\)HS [IQR] | 33 [13–67] | < 0.0001 | 32 [14–73] | < 0.0001 | 57 [25–107] | < 0.0001 |