Table 3 Performance comparison of three indicator screening algorithms represented by the XGBoost algorithm
From: Construction of the XGBoost model for early lung cancer prediction based on metabolic indices
Algorithm | Accuracy (%) | Precision | F1 score | Recall |
|---|---|---|---|---|
Original | 73.73 | 0.77 | 0.77 | 0.78 |
Boruta | 72.55 | 0.74 | 0.77 | 0.8 |
Stepwise regression | 75.29 | 0.76 | 0.79 | 0.83 |