Coronary Artery Disease Detection using Machine Learning

Cardiovascular Diseases (CVD) remain the leading cause of death worldwide, posing a significant challenge to global health systems. Early identification ofi ndividuals a tr isk is crucial for effective prevention and management, especially as sedentary lifestyles and metabolic disorders become increasingly prevalent. In this context, there is a growing need for accessible, reliable, and interpretable tools that can empower individuals and clinicians to assess cardiovascular risk proactively. The objective of this paper is to develop HeartWise Score, a web-based system that leverages machine learning to provide personalized cardiovascular risk predictions using routinely available health and lifestyle information. The methodology for this paper centers on the comparative analysis and deployment of multiple machine learning models-including Artificial N eural Network (ANN), Random Forest, Support Vector Machine (SVM), and Logistic Regression-using a large-scale CVD dataset sourced from Kaggle, comprising 7 0,000 records. After extensive evaluation, Logistic Regression was selected as the final predictive model due to its superior accuracy and interpretability. The dataset was carefully preprocessed, with clinical and engineered features such as BMI, pulse pressure, and mean arterial pressure calculated to enhance predictive performance. The frontend of the application was developed using React and TypeScript, styled with Tailwind CSS to ensure a modern and responsive user experience. The backend was implemented with Flask, which serves predictions from the trained and serialized Logistic Regression model. Key health metrics are computed client-side, and user data is securely transmitted to the backend for real-time cardiovascular risk assessment. The results are presented in an interpretable format, clearly highlighting individual risk factors and providing actionable insights. This work demonstrates the effective integration of machine learning and web technologies to deliver a practical, accessible, and privacy-conscious tool for preventive cardiovascular healthcare.