表 3. Performance of the four models on training set and test set respectively.
四种模型分别在训练集和测试集上的性能
| 分类器 | 准确率 | 精确率 | 召回率 | F1得分 | AUC |
| 训练集结果 | |||||
| LR | 0.807 | 0.648 | 0.760 | 0.699 | 0.904 |
| SVM | 0.843 | 0.727 | 0.747 | 0.737 | 0.891 |
| LightGBM | 0.882 | 0.785 | 0.827 | 0.805 | 0.945 |
| XGBoost | 0.898 | 0.836 | 0.813 | 0.824 | 0.959 |
| 测试集结果 | |||||
| LR | 0.765 | 0.813 | 0.650 | 0.722 | 0.851 |
| SVM | 0.776 | 0.862 | 0.625 | 0.725 | 0.857 |
| lightGBM | 0.800 | 0.829 | 0.725 | 0.773 | 0.906 |
| XGBoost | 0.776 | 0.800 | 0.700 | 0.747 | 0.889 |