Detection of Sepsis Using Machine Learning: A Comprehensive Evaluation of Model Performance

Sepsis is a life-threatening condition arising from the body's response to infection with high mortality and escalating race, representing one of healthcare's remaining challenges. Prediction of sepsis at an early stage leads to patient safety and also helps reduce the cost of healthcare. This paper is focused on investigating the performance of different machine learning (ML) based models for early sepsis detection, given a dataset from Computing in Cardiology Challenge. The data was preprocessed, normalized, and features were selected for model training. Models including Decision Tree, Random Forest, Extra Trees, XGBoost. Results indicate that the Extra Trees classifier provided the highest accuracy, precision and recall (99.7%, 94.0% and 90.0% respectively). The results of this study highlight the potential of ML techniques to improve early sepsis diagnosis, which can result in prompt medical attention and improved patient outcomes.