Table 4 The summary of primary discriminant screening models for different severities of Covid-19 infection
| Â | Discriminant models for disease progression | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Â | Training set with cross-validation | Test set with cross-validation | ||||||||
| Â | Logistic regression | Linear discriminant analysis | K-nearest neighbor | Support vector machine | Decision tree | Logistic regression | Linear discriminant analysis | K-nearest neighbor | Support vector machine | Decision tree |
AUC | 82.9 | 60.9 | 47.6 | 26.3 | 43.4 | 81.9 | 62.3 | 46.0 | 39.0 | 36.9 |
Sensitivity, % | 77.4 | 52.0 | 40.3 | 36.5 | 49.1 | 75 | 50.4 | 49.8 | 40.1 | 41.9 |
Specificity, % | 88.4 | 69.9 | 34.4 | 24.4 | 30.4 | 88.9 | 60.2 | 35.1 | 35.2 | 29.8 |
Positive predictive value, % | 75 | 46.7 | 45.6 | 45.3 | 51.2 | 75 | 59.1 | 55.5 | 33.0 | 30.5 |
Negative predictive value, % | 89.7 | 55.9 | 40.3 | 30.4 | 37.4 | 88.9 | 58.3 | 40.1 | 50.3 | 40.3 |
Accuracy | – | – | – | – | – | 84.6 | 55.4 | 35.4 | 41.3 | 34.5 |
Recall rate | – | – | – | – | – | 75 | 60.0 | 40.1 | 29.9 | 35.3 |