Prediction of Heart Disease Risk in Early Ages with Boosting Techniques

Heart disease is one of the top causes of mortality globally. Early detection can save a person's life. In previous five years, it is found that young individuals are also experiencing heart disease. In order to automate the medical industry and to forecast cardiac illness, artificial intelligence (AI) provides a prediction model. We could employ a number of AI algorithms to predict the risk of heart disease as in previous work, done in this field; generally, authors have used single classifiers. In this work, we have used two approaches of AI. In the first approach, single or base classifiers like logistical regression (LR), support vector machine (SVM) and decision tree (DT) are examined. In the second approach, boosting classifiers such as adaptive boost (AdaBoost), extreme gradient boost (XGBoost) and gradient boost are analyzed, and then comparison is done between single classifiers and boosting classifiers and found that boosting model achieved highest accuracy and the most accurate model as compared with single classifiers in predicting heart disease risk in youngers. So a type of ensemble model called boosting model is employed to prevent overfitting and to offer the highest accuracy as compared to single models. With the XGBoost classifier, we obtained 96.29% accuracy, with AdaBoost classifier, we obtained 91.35% accuracy, and with the gradient boost classifier, we achieved the highest accuracy of 98.76%, which is best in prediction of heart disease ages between 28 and 50.