Prediction of Cardiovascular Disease based on Lipid Profiles

Cardiovascular disease (CVD) remains a global health concern, necessitating continuous efforts to develop predictive models for early risk assessment and intervention. This research employs machine learning and deep learning techniques to compare and predict CVD based on lipid profiles and extends its focus to estimate the future risk of CVD within the studied population. The study encompasses an array of features, including LDL, HDL, total cholesterol, triglycerides, and additional factors like age, BMI, blood glucose, and blood pressure. This study aims to compare the predictive performance of distinct models, logistic regression, SVM and neural networks in assessing the risk of CVD. The initial model, using logistic regression, neural networks and SVM classifier, achieved an accuracy of 85%, 88% and 87% respectively. Furthermore, a second model, designed for future likelihood prediction, demonstrated an accuracy of approximately 83% for logistic regression, 84% for neural networks and 85% for SVM classifier. The future risk of CVD is projected at 65.71%, with a 34.29% likelihood of no future CVD. Additionally, the research classifies CVD risk into four distinct categories, namely atherosclerosis, hyperlipidemia, hypoalphalipoproteinemia, offering valuable insights for tailored prevention and intervention strategies. This work contributes to the ongoing efforts to combat CVD by enhancing predictive accuracy and providing nuanced risk categorization.