Stacking Ensemble Machine Learning Algorithm with an Application to Heart Disease Prediction

Abstract

Mathematics and statistics have a significant impact on the advancement of most trending sciences like machine learning, artificial intelligence, and data science. In this article, we use the Stacking Ensemble Machine Learning Algorithm (SEMLA) to predict heart disease, considering accuracy (acc), diagnostic odds ratio (Dor), F1_score, Matthews correlation coefficient (Mcc), receiver operating characteristics-area under curve (roc-auc), and logloss (log_loss). The data is analyzed using classification learning techniques. We have considered sex, age, cholesterol, fasting blood sugar, the highest rate of heartbeat, type of chest pain, resting electrocardiogram (ECG), angina, depression induced by exercise, peak exercise measurement, major vessel number, a disorder in the blood, and a target attribute to represent the presence and absence of disorders. The approach used allows for the prediction of heart disease and the management of worst-case scenarios. In comparison with the existing models, our proposed model has outperformed other models with an accuracy of 97.28%.