Table 2.
Model | AUC | ACC (%) | Best cutoff | Youden index (%) | Sensitivity (%) | Specificity (%) | F1 score | PPV (%) | NPV (%) |
---|---|---|---|---|---|---|---|---|---|
CatBoost | 0.83 | 75 | 0.195 | 50 | 75 | 75 | 0.56 | 44 | 92 |
GBDT | 0.82 | 71 | 0.16 | 48 | 79 | 69 | 0.53 | 40 | 93 |
LightGBM | 0.82 | 74 | 0.183 | 49 | 75 | 74 | 0.55 | 43 | 92 |
AdaBoost | 0.82 | 79 | 0.494 | 48 | 65 | 83 | 0.57 | 51 | 90 |
Random Forest | 0.82 | 78 | 0.28 | 47 | 66 | 81 | 0.55 | 48 | 90 |
XGBoost | 0.81 | 77 | 0.204 | 47 | 68 | 79 | 0.55 | 46 | 90 |
KNN | 0.8 | 72 | 0.176 | 45 | 73 | 72 | 0.52 | 41 | 91 |
MLP | 0.79 | 73 | 0.162 | 43 | 70 | 73 | 0.52 | 41 | 90 |
LR | 0.79 | 73 | 0.201 | 44 | 71 | 74 | 0.52 | 41 | 90 |
NaiveBayes | 0.76 | 68 | 0.092 | 41 | 74 | 67 | 0.49 | 37 | 91 |
SVM | 0.76 | 74 | 0.149 | 45 | 69 | 75 | 0.53 | 43 | 90 |
CatBoost, categorical boosting; GBDT, gradient boosting decision tree; LightGBM, light gradient boosting; AdaBoost, adaptive boosting; XGBooST, extremely gradient boosting; KNN, K-nearest neighbor; MLP, multilayer perceptron; LR, logistic regression. SVM, support vector machine; ACC, accuracy, PPV, positive predictive value; NPV, negative predictive value.