An Ensemble Machine Learning Method for the Prediction of Heart Disease
Sohaib Asif, Wenhui Yi, Yi Tao, Jinhai Si, Hou Jin
Abstract
Heart disease is a major health issue causing public concern worldwide. Heart disease cases are increasing at a rapid rate every day, so it is very important to predict any such diseases in advance. Therefore, the early diagnosis and prediction of heart disease play a vital role in the correct treatment of patients. In this research article, we have proposed a novel ensemble method using majority voting scheme. First, we compared the performance of different state-of-the-art machine learning classification algorithms for the prediction of heart disease. Six algorithms named as K-nearest neighbor (KNN), Random forest, Naïve bayes, Support vector machine (SVM), XGBoost (XGB) and logistic regression were applied and a comparative study was drawn. Several evaluation techniques were used to evaluate the performance of each algorithm using the Cleveland dataset of the UCI repository of heart patients. We proposed a new ensemble classification model by choosing three algorithms based on the best performance. The proposed ensemble approach yields the highest accuracy, precision, recall and F1 score with 92%, 91.1%, 94%, and 93% respectively on the UCI heart disease dataset. Statistical results demonstrate that the robustness of the ensemble method accurately and reliably distinguishes between heart disease and healthy patients.