Table 3 Performance Comparison for TBI Status Prediction
Method | Accuracy(95% CI) | AUC(95% CI) | PR-AUC(95% CI) | Sensitivity(95% CI) | Specificity(95% CI) |
|---|---|---|---|---|---|
DNN\(^{a}\) | 0.915(0.899,0.931) | 0.781(0.703,0.859) | 0.972(0.954,0.990) | 0.034(0.000,0.171) | 1.000(0.994,1.000) |
PermFIT-DNN\(^{b}\) | 0.915(0.901,0.869) | 0.791(0.713,0.869) | 0.973(0.955,0.991) | 0.048(0.000,0.129) | 1.000(0.994,1.000) |
RF\(^{a}\) | 0.911(0.895,0.927) | 0.653(0.590,0.716) | 0.949(0.933,0.965) | 0.041(0.000,0.159) | 0.996(0.991,1.000) |
PermFIT-RF\(^{b}\) | 0.915(0.909,0.921) | 0.411(0.354,0.468) | 0.878(0.860,0.896) | 0.004(0.000,0.055) | 1.000(0.996,1.000) |
XGBoost\(^{a}\) | 0.899(0.881,0.917) | 0.760(0.684,0.836) | 0.970(0.952,0.988) | 0.107(0.000,0.260) | 0.973(0.967,0.979) |
PermFIT-XGB\(^{b}\) | 0.910(0.894,0.926) | 0.753(0.679,0.827) | 0.968(0.946,0.990) | 0.107(0.000,0.225) | 0.985(0.975,0.995) |
SVM\(^{a}\) | 0.914(0.900,0.928) | 0.798(0.718,0.878) | 0.973(0.955,0.991) | 0.016(0.000,0.151) | 0.998(0.994,1.000) |
PermFIT-SVM\(^{b}\) | 0.915(0.911,0.919) | 0.720(0.655,0.785) | 0.957(0.945,0.969) | 0.004(0.000,0.022) | 1.000(0.998,1.000) |