Table 2.
The ROC Curve Analyses for Predicting Fracture Risk in Each ML-Based Model
| Model | Training Set | Testing Set | ||||
|---|---|---|---|---|---|---|
| AUC Mean | AUC 95% CI | Variables& | AUC Mean | AUC 95% CI | Variables& | |
| RFC | 0.878 | 0.827–0.929 | 7 | 0.872 | 0.819–0.925 | 7 |
| SVM | 0.768 | 0.717–0.819 | 8 | 0.771 | 0.718–0.824 | 8 |
| DT | 0.746 | 0.695–0.797 | 4 | 0.753 | 0.700–0.806 | 4 |
| ANN | 0.819 | 0.768–0.870 | 11 | 0.824 | 0.771–0.877 | 11 |
| XGboost | 0.791 | 0.740–0.842 | 7 | 0.784 | 0.731–0.837 | 7 |
| PMOF | 0.725 | 0.674–0.776 | 1 | 0.776 | 0.723–0.829 | 1 |
| PHF | 0.695 | 0.644–0.746 | 1 | 0.713 | 0.660–0.767 | 1 |
Note: &Variables included in the model.
Abbreviations: RFC, random forest classifier; SVM, support vector machine; DT, decision tree; ANN, artificial neural network; XGboost, eXtreme gradient boosting; AUC, area under curve; 95% CI, 95% confidence interval.