Home // ALLDATA 2022, The Eighth International Conference on Big Data, Small Data, Linked Data and Open Data // View article

Machine Learning Stacking Ensemble Model for Predicting Heart Attacks

Authors:
Muath Obaidat
Alex Alexandrou
Samantha Sanacore

Keywords: machine learning; naïve bayes; random forest; extreme gradient boosting; ensemble; heartattack; accuracy.

Abstract:
To mitigate the extent of one of the world’s leading causes of death, heart attacks, there needs to be an improvement in the technological aspect to predict this disease more accurately. Machine learning methods have come very far in increasing prediction accuracy based on patient data. Ensemble methods have exhibited improvement compared to individual classifier models. For this study, the goal is to develop a Machine Learning model to reach a very high level of accuracy for predicting myocardial infarction, otherwise known as a heart attack. A stacked ensemble model is used in this study and combines a group of three base-level classifiers such as Naïve Bayes, Random Forest, and Extreme Gradient Boosting (XGBoost). This model will help identify those who are at risk and prevent heart attacks, therefore, lowering the mortality rate globally. Diversity among strong classifiers used in this model will be a more effective way to achieve the highest accuracy. The metrics used to evaluate the prediction performances are accuracy, Area Under the Curve (AUC), specificity, precision, and sensitivity. This process is carried out using RStudio and the results indicate that the proposed stacked ensemble method had a better performance under every evaluation metric compared to the individual base-level classifiers that were utilized.

Pages: 8 to 14

Copyright: Copyright (c) IARIA, 2022

Publication date: April 24, 2022

Published in: conference

ISSN: 2519-8386

ISBN: 978-1-61208-945-4

Location: Barcelona, Spain

Dates: from April 24, 2022 to April 28, 2022