Table 4 Performance of the two models, evaluated on an external dataset
From: Monitoring mandibular movements to detect Cheyne-Stokes Breathing
Metrics | Scale (Worst-Best) | Logistic model | Decision tree model |
|---|---|---|---|
ROC-AUC | 0.5–1 | 0.93 | 0.92 |
Balanced Accuracy | 0–1 | 0.92 | 0.92 |
True positive rate (Sensitivity) | 0–1 | 0.92 | 0.92 |
True negative rate (Specificity) | 0–1 | 0.92 | 0.92 |
Positive predictive value | 0–1 | 0.93 | 0.93 |
Negative predictive value | 0–1 | 0.91 | 0.91 |
False negative rate | 1-0 | 0.08 | 0.08 |
False positive rate | 1-0 | 0.08 | 0.08 |
Balance error | 1-0 | 0.08 | 0.08 |
Mean misclassification error index | 1-0 | 0.08 | 0.08 |