Table 1 Results based on three models with and without differential features.
From: Machine learning-aided risk prediction for metabolic syndrome based on 3 years study
Model | Threshold | AUC | Accuracy | Precision | Recall | F1-score | Specificity | F2-score |
|---|---|---|---|---|---|---|---|---|
Without DNFs and DSFs | ||||||||
XGBoost | 0.147 | 0.918 ± 0.003 | 0.833 | 0.40 | 0.85 | 0.55 | 0.83 | 0.69 |
Stacking | 0.116 | 0.917 ± 0.003 | 0.812 | 0.37 | 0.88 | 0.52 | 0.80 | 0.69 |
Random Forest | 0.156 | 0.908 ± 0.003 | 0.804 | 0.36 | 0.88 | 0.51 | 0.79 | 0.68 |
With DNFs and DSFs | ||||||||
XGBoost | 0.144 | 0.930 ± 0.002 | 0.849 | 0.43 | 0.87 | 0.58 | 0.85 | 0.72 |
Stacking | 0.125 | 0.928 ± 0.002 | 0.837 | 0.41 | 0.89 | 0.56 | 0.83 | 0.72 |
Random Forest | 0.177 | 0.916 ± 0.002 | 0.825 | 0.39 | 0.87 | 0.54 | 0.82 | 0.70 |
