Table 1 The precision and recall of our approach and the three machine learning methods at 0 to 5-days prior to AKI events
Prediction Time | Our approach (Precision/Recall) | Logistic Regression (Precision/Recall) | Random Forest (Precision/Recall) | AdaBoostM1 (Precision/Recall) |
|---|---|---|---|---|
AKI Time-0 | 0.826/0.93 | 0.779/0.781 | 0.749/0.742 | 0.759/0.765 |
AKI Time-1 | 0.713/0.821 | 0.716/0.757 | 0.687/0.686 | 0.713/0.734 |
AKI Time-2 | 0.64/0.766 | 0.652/0.646 | 0.631/0.582 | 0.639/0.688 |
AKI Time-3 | 0.651/0.664 | 0.633/0.609 | 0.611/0.567 | 0.628/0.615 |
AKI Time-4 | 0.654/0.668 | 0.631/0.617 | 0.603/0.578 | 0.621/0.651 |
AKI Time-5 | 0.596/0.708 | 0.618/0.602 | 0.578/0.565 | 0.606/0.642 |