Machine Learning–Driven Ontology System for Coronary Heart Disease Diagnosis

Abstract

Coronary heart disease is a disease that affects the larger coronary arteries on the surface of the heart and has been seen to be one of the major factors contributing to morbidity and death globally. This project presents a machine learning-driven ontology system for diagnosing coronary heart disease (CHD) based on patient data and clinical parameters. The system integrates a trained machine learning model, developed with the Framingham Heart Study dataset on the Weka platform, to predict the risk of CHD risk by extracting classification rules and embedding them within an ontology framework built in Protégé. The ontology models patient demographics, symptoms, and risk factors, providing a structured and semantically rich representation to support accurate diagnosis and clinical decision-making. The decision tree machine learning model achieved an 82.8 percent accuracy, and the ontology model was evaluated to be coherent and consistent. The system was further deployed to the web, enabling user-friendly interaction. By combining machine learning with ontology, this project not only leverages data-driven predictive power but also enhances interpretability, transparency, and adaptability in diagnosis. This synergy bridges the gap between statistical prediction and structured medical knowledge, ultimately contributing to more effective CHD risk assessment and improved patient outcomes in clinical practice.