Prediction of Heart Disease Using Machine Learning: A Systematic Literature Review

This article aims to carry out a systematic review of the research works that deal with the topic of Machine learning (ML) and Deep learning (DL) to predict heart disease. For this purpose, an exhaustive search was carried out and after the analysis of the documentation collected, aspects such: countries have more studies been done on use of ML and DL, techniques most used and that had the best accuracy, tools, metrics, kind of heart disease and variable selection algorithms, which served as the basis for the elaboration of this document. The intent is to contribute to more profound understanding of methodologies, techniques and metrics with the applications of machine learning for predict heart disease. The results of the study showed that India, China and Pakistan were the countries with most studies on the use of ML and DL to predict heart disease, also Random Forest, SVM and Logistic Regression were the most used techniques, of which XGBoost, Ensemble Deep learning and Stacking were the ones that obtained the best accuracy results. Python was the tool considered the best. The most concurrent metrics used was Accuracy, Precision and F1-Score, the type of disease that has been applied was Coronary Artery Heart, as the selection algorithms were Kernel and Information Gain. Due to the paucity of this studies on heart disease using machine learning and deep learning, this work also points the way to new research.