Early prediction of sepsis in the ICU: a comparative analysis of multiple machine-learning algorithms using the MIMIC-III database - PubMed
4 hours ago
- #ICU
- #sepsis-prediction
- #machine-learning
- Machine-learning models were developed to predict sepsis onset beyond the first 24 hours of ICU admission using the MIMIC-III database.
- Nine algorithms were evaluated, with XGBoost-DART achieving the highest AUROC (0.881) and best performance in accuracy, F1-score, and specificity.
- The XGBoost-DART model demonstrated strong clinical utility through decision-curve analysis, enabling timely identification of high-risk patients.