Heart-related Clinical Biomarker Classification through Machine Learning Algorithms

Abstract

Abstract Views: 0

Heart diseases continue to be a major global cause of morbidity and death. Their timely and accurate diagnosis for improved patient outcomes is direly needed. Clinical biomarkers for the timely diagnosis of heart diseases are known but underutilized due to the use of conventional analytical methods that lower the efficiency to handle large datasets. Furthermore, conventional methods also fail to incorporate demographic biomarkers such as age and hemodynamic biomarkers such as heart rate and diastolic blood pressure. This significantly influences heart diseases. Using cutting edge machine learning (ML) techniques including Lasso regularization, support vector machine (SVM), and gradient boosting (GB), this study investigated the importance of clinical biomarkers for heart disease prediction. Troponin and Creatinine Kinase - MB (CK-MB) were found to be the most significant predictors among the examined characteristics in every model, underscoring their crucial importance in the diagnosis of myocardial ischemia and damage. Diastolic blood pressure was also found to be an adequate predictor, highlighting its role in increasing cardiovascular risk because of autonomous dysfunction. While SVM and GB performed strongly in managing intricate data relationships, Lasso regularization successfully decreased feature redundancy. The results support the use of clinically applicable biomarkers in conjunction with machine learning to improve the accuracy of diagnosis and also opens the door to the personalized treatment of heart disease. Validating these findings in a variety of populations and adding more biomarkers for a thorough risk assessment should be the main goals of future studies.