Predictive Models for Suicide Attempts in Major Depressive Disorder and the Contribution of EPHX2: A Pilot Integrative Machine Learning Study
Table 2
Classification performance of MDD recognition.
AUC (95% CI)
Sensitivity (95% CI)
Specificity (95% CI)
Accuracy (95% CI)
PPV (95% CI)
NPV (95% CI)
-score
D-Model A
0.901 (0.850, 0.951)
91.7% (84.9%, 96.2%)
63.6% (45.1%, 79.6%)
85.2% (78.3%, 96%)
89.3% (82.0%, 94.3%)
70.0% (50.6%, 85.3%)
0.850
D-Model B
0.938 (0.898, 0.977)
93.6% (87.2%, 97.4%)
75.8% (57.7%, 88.9%)
89.4% (83.2%, 94.0%)
92.7% (86.2%, 96.8%)
78.1% (60.0%, 90.7%)
0.894
D-Model C
0.928 (0.886, 0.969)
91.7% (84.9%, 96.2%)
54.5% (36.4%, 71.9%)
83.1% (75.9%, 88.9%)
87.0% (79.4%, 92.5%)
66.7% (46.0%, 83.5%)
0.825
AUC: area under the receiver operating characteristic curve; PPV: positive predictive value; NPV: negative predictive value; D-Model: model for recognising major depressive disorder.
