Abstract :

Hospital readmissions place a significant burden on healthcare systems and often indicate gaps in patient care. Predicting patient readmission risk using Electronic Health Records (EHR) can improve clinical decision-making and resource management. This project proposes a machine learning–based system to analyze patient medical history, diagnoses, treatments, and demographic data. The system extracts meaningful patterns from structured EHR data to identify patients at high risk of readmission. Advanced preprocessing techniques handle missing values and class imbalance. Multiple predictive models are evaluated to achieve high accuracy and reliability. The system enables early intervention and personalized care planning. It assists healthcare providers in reducing avoidable readmissions. The proposed approach improves efficiency and patient outcomes. Overall, the system offers a data-driven solution for predictive healthcare analytics.