Prognosis of Cardiovascular Disease Using Machine Learning Procedures

Abstract

The topmost crucial muscular body part is our heart, as it pumps blood to all the other organs in our body. Being the essential organ, mortals suffering from heart disease over the last twenty years has been the deadliest disease globally and top number one destroyer for human. Over the recent years, the health industry and technology have worked together to find ways to cut back the risk of cardiac diseases in humans. For early disease prediction, machine learning is a necessity for healthcare as it functions without human interaction. In this paper, a cardiovascular data set with 70,000 data and 12 attributes are analyzed and implemented for the early prognosis of cardiovascular disease. Using the voting ensemble classifier, we combined five different machine learning algorithms to achieve good overall accuracy. K - nearest neighbor classifier gained an ac-curacy of 75%, which was the best amongst Logistic Regression, Random Forest, Gradient Boosting, and Bernoulli Naive Bayes. This proposal benefits and eases the work for clinicians and doctors and provides appropriate care for heart disease patients.