Analysis of Machine Learning Algorithms in Prediction of Cardiovascular Diseases

Abstract

Heart failure is considered as one among the most fatal diseases in the contemporary world. Diabetes mellitus, hypertension, and dyslipidemia are considered as the observed predictors of cardiovascular disease. Few routine style risk factors include depression, physical inactivity, smoking, alcohol consumption, stress, food habits and obesity which are the major causes for cardiovascular disease. In India, heart failure among people is increasing at an alarming rate because there is lack of proper estimation for the root cause of cardiovascular diseases and the absence of surveillance programme in order to track the occurrence, extensiveness and outcomes of heart failure. Data mining techniques prove to be an efficient approach in predicting the risk of cardiovascular diseases in the data deluge age. In this research study, data mining techniques are applied to get useful information from medical reports of patients. Using machine learning algorithms, the impact of each risk factor on heart disease is predicted. Firstly, the heart disease dataset is collected from the Cleveland Heart Disease database. With the help of the dataset, the attributes significant to the heart attack prediction are extracted. The dataset is split into training and test dataset. Different classification techniques are applied on preprocessed data to measure their accuracy in predicting the risk of heart disease. Two such algorithms are Logistic Regression and Gradient Boosting Algorithm. The objective is to attain high accuracy in the prediction of risk of cardiovascular diseases among patients. In order to prevent the occurrence of the cardiovascular diseases, the prevalence of risk factors should be minimized. Further, early conclusion and treatment can enhance quality and future of individuals who have heart disappointment.