Table 5.
Performance comparison with conventional logistic regression model for total set (maximum sensitivity criterion).
| AUC | 95% Confidence Interval | p for Comparison with CLR | |
|---|---|---|---|
| CLR | 0.568 | 0.563–0.572 | Reference |
| NBC | 0.669 | 0.663–0.675 | <0.001 |
| RFC | 0.775 | 0.770–0.780 | <0.001 |
| DAC | 0.661 | 0.655–0.667 | <0.001 |
| KNNC | 0.8723 | 0.868–0.877 | <0.001 |
| SVMC | 0.515 | 0.509–0.522 | <0.001 |
| DTC | 0.537 | 0.534–0.541 | <0.001 |
CLR: conventional logistic regression; NBC: naïve Bayes classification; RFC: random forest classification; DAC: discriminant analysis classification; KNNC: K-nearest neighbor classification; SVMC: support vector machine classification; DTC: decision tree classification; and AUC: area under the curve.