Table 2 Performance of machine learning classifiers in predicting recurrence, types of recurrence, and distant metastasis locations for validation cohort.
Classifier | Accuracy (%) | Recall (%) | F1-Score (%) |
|---|---|---|---|
Recurrence | |||
LGBM | 84 | 89 | 89 |
XGBoost | 82 | 86 | 87 |
NN | 83 | 83 | 87 |
SVM | 80 | 78 | 84 |
RF | 83 | 94 | 89 |
KNN | 74 | 75 | 80 |
Types of recurrence | |||
LGBM | 86.75 | 87 | 89 |
XGBoost | 86.70 | 87 | 89 |
NN | 86.77 | 87 | 89 |
SVM | 86.71 | 87 | 89 |
RF | 86.72 | 87 | 89 |
KNN | 86.36 | 86 | 89 |
Distant metastasis locations (Bone Vs. Others) | |||
LGBM | 74 | 70 | 70 |
XGBoost | 74 | 70 | 69 |
SVM | 69 | 69 | 69 |
RF | 74 | 69 | 69 |
