Review of Machine Learning Based Methodologies for Detection of Heart Diseases

Abstract

Nowadays, the number of patients having heart disease has significantly increased. This is due to our lifestyles, globalization, industrialization, and a variety of other things. This paper presents the review of various machine learning-based methodologies to diagnose heart disease by analyzing the heart sounds. In a later section of the paper, we discussed various machine learning methodologies and their insight to detect the abnormality in the heart through analysis of acoustic sounds. Most of the work has been carried on the segmentation of heart sounds and their accuracy mainly depends on accuracy of segmentation methods and not on the content of heart sounds. Therefore, it is very important to propose classification methods over segmentation and here the role of machine learning techniques comes into the focus. We mainly emphasize on identifying the positive outcomes of different research and shortcomings in the existing literature. Based on this, we will discuss the next steps in tackling this study topic.