Table 4 Final accuracy estimates of the BNN and MNN models for hair greying prediction designated in a 849-sample cohort using 10-fold cross-validation procedure
Model | Â | AUC | Sensitivity | Specificity | PPV | NPV |
|---|---|---|---|---|---|---|
12-variable BNN | greying vs. no greying | 0.873 | 0.734 (243/331) | 0.853 (442/518) | 0.762 (243/319) | 0.834 (442/530) |
14-variable MNN | no greying | 0.864 | 0.886 (459/518) | 0.643 (196/305) | 0.808 (459/568) | 0.769 (196/255) |
mild greying | 0.791 | 0.589 (149/253) | 0.821 (468/570) | 0.594 (149/251) | 0.818 (468/572) | |
severe greying | 0.875 | 0.077 (4/52) | 0.997 (769/771) | 0.667 (2/6) | 0.941 (769/817) |