Predicting Hospital Readmission Risk Among Diabetic Patients

This project analyzes the readmission patterns of diabetic patients at HealthFirst Multispeciality Hospital. Using a dataset from the UC Irvine Machine Learning Repository, it leverages data cleaning, Principal Component Analysis (PCA), Random Forest, K-Means clustering, and logistic regression to predict readmission risk. While Random Forest achieved moderate accuracy (48.25%), logistic regression underperformed due to data imbalance. Key insights revealed that treatment complexity, hospitalization history, age-related risks, and emergency visits significantly influence readmission. Actionable strategies include optimizing medication regimens, enhancing telehealth monitoring, and leveraging data analytics to proactively identify high-risk patients, improving patient care and hospital efficiency.