Table 3.
Each evaluation index of Cli_Pat_Rad_PETCT model in five machine learning algorithms.
| Model name | AUC | 95% CI | Accuracy | Sensitivity | Specificity | PPV | NPV | |
|---|---|---|---|---|---|---|---|---|
| SVM | 0.923 | 0.8704–0.9762 | 0.894 | 1.000 | 0.787 | 0.825 | 1.000 | Train |
| SVM | 0.857 | 0.7473–0.9664 | 0.896 | 1.000 | 0.826 | 0.828 | 1.000 | Test |
| KNN | 0.905 | 0.8504–0.9603 | 0.830 | 0.894 | 0.766 | 0.792 | 0.878 | Train |
| KNN | 0.724 | 0.6762–0.8723 | 0.792 | 1.000 | 0.636 | 0.706 | 0.878 | Test |
| LR | 0.828 | 0.7442–0.9131 | 0.787 | 0.830 | 0.745 | 0.765 | 0.814 | Train |
| LR | 0.807 | 0.6711–0.9420 | 0.833 | 0.958 | 0.708 | 0.767 | 0.944 | Test |
| LightGBM | 0.875 | 0.8062–0.9443 | 0.819 | 0.830 | 0.809 | 0.812 | 0.826 | Train |
| LightGBM | 0.801 | 0.6673–0.9354 | 0.833 | 0.833 | 0.870 | 0.833 | 0.833 | Test |
| NaiveBayes | 0.807 | 0.7223–0.8934 | 0.766 | 0.809 | 0.723 | 0.745 | 0.826 | Train |
| NaiveBayes | 0.776 | 0.6312–0.9213 | 0.833 | 0.792 | 0.875 | 0.864 | 0.833 | Test |
AUC, area under the curve; CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value; SVM, support vector machine; KNN, K-nearest neighbors; LR logistic regression; LightGBM, light gradient boosting machine.