SMOTE-TOMEK: A Hybrid Sampling-Based Ensemble Learning Approach for Sepsis Prediction

Sepsis is a lethal condition that requires early detection and intervention to improve patient outcomes. Machine learning algorithms have shown promise in detecting sepsis using electronic health records. This Study evaluates the performance of an ensemble based machine learning model for prediction of sepsis earlier using six vital signs. The research validates the tool and compares it to existing methodologies. The results suggest that the technique can accurately forecast sepsis up to 48 hours in advance. The method had high sensitivity and specificity, and it was resistant to missing data. Early prediction is critical for improving outcomes of sepsis patients, and delay in treatment increases the risk of patients. Therefore, machine learning algorithms can play a crucial role in early sepsis detection and prediction, leading to better patient outcomes.