MACHINE LEARNING FOR EARLY HEART DISEASE DETECTION

<p>Heart disease has been proven to be the leading cause of death for both men and women, killing about 697,000 <br>people in the US in 2020, which makes it a major concern to be dealt with. Identifying heart disease in patients <br>could be quite challenging due to several contributory risk factors, which requires some high level of <br>techniques. Machine learning proves to be effective in predicting heart disease in patients given the <br>contributory risk factors. In this study, we used seven different machine learning algorithms such as K-Nearest <br>Neighbors, support vector machine, Decision Tree, Random Forest, Logistic Regression, Neural Network, and <br>AdaBoost, to predict whether a person has heart disease using the Cleveland Heart Disease Dtaset from the UCI <br>Repository. Our goal was to identify the most accurate algorithm for this task. Linear support vector machine <br>(SVM) and Elastic net logistic regression have the best overall performance with an overall performance of <br>89%. These findings highlight the potential of machine learning in effectively improving heart disease diagnosis <br>and management, offering promising avenues for improving patient care and outcomes.</p>