Heart Disease Prediction Using Machine Learning: A Systematic Literature Review

Heart disease, categorized as a cardiovascular condition, stands as a prominent factor in worldwide mortality, accounting for approximately 32% of all deaths globally. It manifests when the buildup of arterial plaque obstructs the circulation of blood to the heart or brain, potentially resulting in a stroke or heart attack. Early identification of heart disease is needed to reduce mortality rates and improve decision-making in the prevention and treatment of high-risk individuals. This can be done by using a prediction model. To offer a comprehensive examination of machine learning research about the prediction of cardiac disease, a systematic literature review (SLR) was undertaken. Based on these papers, researchers can focus on four main research questions. The UCI dataset was commonly used in 25 out of 32 papers. Random forest was the most popular machine learning algorithm, used in six papers. The limitations identified by the authors mainly revolve around the dataset, including the need for a larger one. Other limitations involve the inability to detect subtle changes and the use of oversimplified algorithms for prediction. To address these limitations, future research can explore new and larger datasets, experiment with different algorithms, and consider advancements in software and hardware.