Predictive Modeling of Coronary Heart Disease Incidence Using Machine Learning: A Comprehensive Epidemiological Analysis with Web-Based Implementation

Coronary Heart Disease (CHD) remains the leading cause of mortality worldwide, accounting for approximately 17.9 million deaths annually. This paper presents a comprehensive machine learning framework for predicting CHD incidence, integrating epidemiological trend analysis with practical clinical deployment through a web-based application. We developed and evaluated multiple predictive models including Logistic Regression, Random Forest, Gradient Boosting, Support Vector Machine, and Neural Networks using the Framingham Heart Study dataset (4,240 patients, 16 clinical features). The Random Forest model achieved superior performance with 85.6% accuracy, 84.2% sensitivity, 86.8% specificity, and an AUC of 0.912, outperforming traditional risk assessment tools. Our epidemiological analysis revealed significant global disparities, with CHD mortality rates ranging from 310 per 100,000 in Eastern Europe to 85 per 100,000 in Western Europe. We deployed the predictive model through an interactive Streamlit web application, enabling real-time risk assessment for clinical decision support. This work demonstrates the feasibility of integrating advanced machine learning with epidemiological insights for personalized cardiovascular risk prediction, providing a scalable solution for early intervention and preventive healthcare strategies.