A Machine Learning Framework for Early Prediction of Chronic Diseases

Chronic diseases, such as Alzheimer’s and cardiovascular conditions, pose significant global health challenges, necessitating early detection to improve outcomes and reduce costs. This study builds a machine learning (ML) framework using the U.S. Chronic Disease Indicators (CDI) and Alzheimer's datasets. It does this by combining a new hybrid feature selection method with advanced classification algorithms. Gradient boosting models (XGBoost, LightGBM) do better than traditional classifiers, and the framework achieves up to 93.2% accuracy and 0.96 AUC-ROC. It improves early detection by 25–30% and makes computations 30% easier. It provides us useful information about risk factors like APOE ε4 and cholesterol levels. These findings support data-driven healthcare policies and preventive strategies, laying a foundation for scalable, AI-driven chronic disease management.