Estimating Ten-Year Coronary Heart Disease Risk through Machine Learning Techniques

This study investigates estimating ten-year Coronary Heart Disease risk (CHD). It found that age and systolic blood pressure play crucial roles in making these predictions. Among various models tested, the Support Vector Machine (SVM) classifier performed best, showing high accuracy and F1 score. Although decision tree demonstrated slightly better accuracy, the SVM outperformed other models in most metrics. Balancing the dataset using SMOTE improved sensitivity. The study suggests that with more data, especially from minority groups, the models could become even more accurate in predicting CHD risk.