Ensemble Method for Heart Disease Prediction

Abstract

The ensemble classifier amalgamates different types of classifier in a manner which increases performance in comparison to other individual classifier. The proposed framework aimed to predict risk of heart disease at an early stage in patients using different classification algorithms in machine learning, from data mining perspective. The objective of the proposed system is to assess the most influential factors of heart disease moreover precisely predicting risk of heart disease at an early stage. Gaussian naive bayes classifier, Decision tree, Random forest & XG Boost classifier was used to predict overall risk of heart disease. The proposed framework uses feature selection technique and ensemble classifier to predict the chances of heart ailment to improve the classification performance as well as reducing training time and over fitting. Additionally, the resulting derived gives high specificity rate which makes an easy to use tool for cardiologists to screen patients who have a Higher possibility of having the heart ailment & transfer those patients for further analysis.