Impact of Machine Learning Algorithms on Heart Disease Datasets

Abstract

Machine learning is advancing with great speed in the healthcare areas by developing algorithms and obtaining information regarding abnormalities in the human body. It has been seen that heart diseases are one of the leading causes of death. Thus, early detection leads to proper treatment and thus leading to saving many lives. Regular monitoring of the heart via different tests help in early detection. Automation of the detection process would not only increase efficiency but also decreases life risk. This paper intends to find out the effect of machine learning algorithms on heart disease prediction. Algorithms such as Logistic Regression, XGBoost, Support Vector Machines, and some others were implemented over different datasets. The performance of each of the algorithms was analysed along with the variations in their hyper parameters for testing. Promising results were obtained for both the datasets with the help of Random Forest, Decision Trees and Neural Networks. The paper also provides the relative feature importance while making the predictions