Advanced Machine-Learning Technologies for Coronary Artery Disease Prediction Using Heterogeneous Data

Epidemiological studies have played an important role in explaining the risk factors associated with cardiovascular diseases (CVD) and identifying opportunities for prevention. Early CVD intervention can prevent the progression of these diseases and effectively lower the mortality rate. In particular, using coronary artery calcium (CAC) scores has been proven to enhance the prediction of coronary heart disease events. Thus, the early prediction of high-risk CAC allows individuals to prevent coronary heart disease (CHD) from progressing to extreme symptoms and illnesses. The process of identifying and using various predictors to predict CAC has received increasing attention, which has helped clinicians and patients make informed decisions. In this discipline, traditional methods rely on statistical approaches to calculate the effects of clinical and demographic factors on CAC score prediction, enabling the identification of relevant features. However, with the advancement of machine learning (ML) techniques, it is now possible to train CAC models that offer highly accurate and stable predictions. This research investigates the application of non-parametric methods to identify and present non-linear relationships within cardiovascular disease datasets derived from the King Faisal Specialist Hospital and Research Centre in Saudi Arabia, without assuming normal distributions. Furthermore, we will focus on efficient feature extraction to handle the dataset effectively. Additionally, we will explore the utility of regularised regression for predicting CAC scores.