Abstract: In critical care settings, early detection of medical emergencies plays a crucial role. Traditional methods often rely on healthcare professionals manually observing patients, which can lead to the oversight of early signs of declining health. This study introduces a real-time monitoring system that utilizes machine learning to analyze essential physiological data and identify potential health risks before they deteriorate. The system focuses on patients with sepsis and monitors key vital signs, including heart rate, oxygen saturation, body temperature, respiratory rate, systolic blood pressure, and diastolic blood pressure. We employed various machine learning algorithms, including Logistic Regression, Decision Tree, and Random Forest, to effectively assess the risk of sepsis. Careful data preprocessing and feature selection significantly enhanced the performance of the models. Among the models tested, the Random Forest classifier achieved the highest accuracy. Additionally, we integrated the trained model into an easy-to-use dashboard built on Streamlit, enabling real-time patient monitoring, anomaly detection, and comprehensive risk analysis. This system is designed to assist healthcare professionals in promptly addressing hidden medical emergencies, thereby improving clinical decision-making. Furthermore, SHapley Additive Explanations (SHAP) analysis was used to explain the model predictions, enhancing transparency and trust.

Keywords: Machine Learning, Sepsis Prediction, Healthcare Monitoring, Random Forest, Data Analytics, Streamlit Dashboard

Downloads: | DOI: 10.17148/IJIREEICE.2026.14542