Table 3 ROC analysis of TRG and pCR classification models
TRG | pCR | |||||||
|---|---|---|---|---|---|---|---|---|
AUC | 95% CI | Sensitivity (%) | Specificity (%) | AUC | 95% CI | Sensitivity (%) | Specificity (%) | |
LASSO | ||||||||
LR | 0.734 | 0.643–0.813 | 48.4 | 91.5 | 0.801 | 0.696–0.882 | 64.3 | 84.6 |
RF | 0.943 | 0.883–0.978 | 90.3 | 92.7 | 0.912 | 0.827–0.964 | 78.6 | 92.3 |
DT | 0.838 | 0.757–0.901 | 87.1 | 80.5 | 0.870 | 0.775–0.935 | 78.6 | 95.4 |
KNN | 0.777 | 0.689–0.850 | 51.6 | 93.9 | 0.945 | 0.870–0.984 | 85.7 | 98.5 |
PCA | ||||||||
LR | 0.761 | 0.697–0.828 | 100 | 0.0 | 0.597 | 0.552–0.642 | 54.0 | 45.0 |
RF | 0.930 | 0.849–1.000 | 90.3 | 40.0 | 0.693 | 0.556–0.830 | 54.0 | 35.0 |
DT | 0.633 | 0.59–0.676 | 73.3 | 0.0 | 0.666 | 0.570–0.762 | 40.0 | 63.3 |
KNN | 0.840 | 0.788–0.892 | 100 | 0.0 | 0.712 | 0.557–0.867 | 78.7 | 56.7 |