Use of Machine Learning Techniques in Healthcare: A Brief Review of Cardiovascular Disease Classification
12 Pages Posted: 27 Aug 2020
Date Written: June 18, 2020
Abstract
One of the major leading cause of fatalities globally, cardiovascular diseases are a growing social concern worldwide. With the advent of technology, medical domain has benefited tremendously with the intersection of machine learning and wearable technology towards providing seamless solutions that are highly accurate, reliable and robust. This benefits the patient community with early detection and reduced costs of medical care; as well as provides efficient, scalable, accurate and reliable prediction systems for the medical fraternity. This paper presents an extensive survey of machine learning algorithms used in prediction/classification of various cardiovascular diseases. We present insights into various modalities such as heart sounds, electronic health records, physiological signals and CT images for successful detection of cardiac disease, and also present highlights of popular machine learning systems, fuzzy systems, hybrid systems. From this review, it is noted that SVM has been popularly used, followed by neural networks and ensemble techniques. Highest accuracy of over 95% were attained by ensemble techniques, followed by SVM and CNN.
Keywords: - Cardiovascular disease detection, Deep learning, Healthcare, Heart disease classification, Machine learning, physiology
JEL Classification: c60,c90
Suggested Citation: Suggested Citation