Abstract:
Cardiovascular diseases (CVD) include a variety of conditions affecting the heart and blood vessels, often arising from complex interplay of genetic predisposition, lifestyle factors, and environmental influences. Myocardial Infarction (MI) stands out as a critical event within the domain of CVD, and is understood to be a strong indicator of progressing CVD. This research begins with a thorough exploration of the challenges inherent in addressing CVD and the intricate nature of cardiovascular health. A critical focus is placed on the 30-day mortality rate as a pivotal metric, emphasizing its significance in understanding the immediate consequences of a cardiovascular event. This project presents a complete approach towards the development of a risk scoring system using advanced machine learning techniques, specifically tailored for assessing cardiovascular risk following the first incidence of myocardial infarction in Indian patients. Traditional machine learning models, along with a novel deep learning-based approach utilizing Graph Convolutional Networks (GCN), are implemented and rigorously evaluated. This project introduces a structured methodology for data processing, model training, and evaluation ensuring the reliability and reproducibility of the results. The experimentation process involves baseline models i.e. traditional machine learning algorithms, and a GCN network. The outcomes are meticulously presented through apt visualizations (where required) and relevant metrics to get a comprehensive understanding of the models’ performance. The findings are interpreted within the context of the research objectives, drawing comparisons with existing literature. The strengths and limitations of the risk scoring system are addressed, with considerations for potential biases and/or improvements. The results and insights gleaned from this research hold promise for improving CVD risk assessment and response methods in India.
