Table 2 Comparison of the classification performance of four models.
From: Enhanced value of chest computed tomography radiomics features in breast density classification
Model | Class | Precision | Sensitivities | F1-scores | Overall accuracy | Kappa |
|---|---|---|---|---|---|---|
① One Vs Rest-Logistic | Class 0 | 0.00 | 0.00 | 0.00 | 0.22 | 0.043 |
Class 1 | 0.22 | 1.00 | 0.37 | |||
Class 2 | 0.00 | 0.00 | 0.00 | |||
Class 3 | 0.00 | 0.00 | 0.00 | |||
② Gradient Boosting | Class 0 | 0.84 | 0.98 | 0.90 | 0.84 | 0.783 |
Class 1 | 0.79 | 0.65 | 0.71 | |||
Class 2 | 0.82 | 0.72 | 0.77 | |||
Class 3 | 0.88 | 0.96 | 0.92 | |||
③ Random Forest | Class 0 | 0.87 | 1.00 | 0.93 | 0.85 | 0.805 |
Class 1 | 0.79 | 0.68 | 0.73 | |||
Class 2 | 0.82 | 0.72 | 0.77 | |||
Class 3 | 0.90 | 0.98 | 0.94 | |||
④ XGBoost | Class 0 | 0.89 | 1.00 | 0.94 | 0.87 | 0.828 |
Class 1 | 0.78 | 0.72 | 0.75 | |||
Class 2 | 0.84 | 0.74 | 0.79 | |||
Class 3 | 0.94 | 0.98 | 0.96 |
