Abstract

The medical field comprises of various techniques. Yet, the Data mining is playing a crucial role in determining the future of medications and patients’ state. This is because of the reliability offered by the various classification techniques. Still, accurate prediction of heart disease is becoming more and more challenging due to the influence of the various factors extracted from patients. Identifying these factors is a crucial research task. In such a scenario, the individual classification algorithms fail to produce perfect models capable of accurately predicting the heart disease. Hence, by introducing the ensemble learning methods, higher performance could be achieved leading to the accurate prediction of heart diseases. In this research work, the performance of the three ensemble classifiers namely Bagging, Stacking and AdaBoost is experimented and evaluated on various folds of cross validation with benchmark dataset for heart disease prediction. The base learners considered for constructing the ensemble are well known classifiers namely Support Vector Machine, Naive Bayes and K-Nearest Neighbour. The results illustrate the improved performance in terms of performance metrics and provide a better understanding of the accuracy, reliability and usefulness of the ensemble models in favouring improved performance for heart disease prediction.