Table 2 Diagnostic performance analysis of LR, DT and GBDT models
Classifier | Group | Sensitivity (95%CI) | Specificity (95%CI) | Accuracy (95%CI) | PPV | NPV | AUC (95%CI) |
|---|---|---|---|---|---|---|---|
LR | Training cohort | 0.918(0.839–0.996) | 0.492(0.400-0.584) | 0.792(0.729–0.855) | 0.426 | 0.933 | 0.815 (0.744–0.885) |
Internal validation cohort | 0.941(0.840-1.000) | 0.437(0.312–0.576) | 0.792(0.744–0.841) | 0.417 | 0.955 | 0.815 (0.602–0.904) | |
External test cohort | 0.852(0.678–0.999) | 0.688(0.573–0.803) | 0.818(0.732–0.964) | 0.424 | 0.956 | 0.910 (0.810–0.978) | |
DT | Training cohort | 0.966(0.914–0.997) | 0.639(0.551–0.727) | 0.870(0.818–0.922) | 0.525 | 0.979 | 0.883 (0.826–0.941) |
Internal validation cohort | 0.941(0.840–0.996) | 0.429(0.290–0.568) | 0.790(0.695–0.885) | 0.414 | 0.944 | 0.803 (0.587–0.845) | |
External test cohort | 0.787(0.586–0.988) | 0.500(0.376–0.625) | 0.727(0.628–0.826) | 0.289 | 0.901 | 0.700 (0.545–0.856) | |
GBDT | Training cohort | 0.986(0.952–0.997) | 0.770(0.693–0.847) | 0.923(0.882–0.964) | 0.639 | 0.993 | 0.981 (0.957-1.000) |
Internal validation cohort | 0.882(0.744-1.000) | 0.714(0.588–0.841) | 0.833(0.746–0.920) | 0.570 | 0.934 | 0.815 (0.704–0.920) | |
External test cohort | 0.918(0.784–0.999) | 0.563(0.442–0.687) | 0.844(0.764–0.925) | 0.352 | 0.964 | 0.819 (0.686–0.952) |