Survey on Analysis, Classification and Early Detection of Heart Diseases using Heartbeat Features and Machine Learning Algorithms

The problems related to heart are mapped by ECG test. Electrocardiogram maps electrical activity of heart that changes with time. ECG test is first test to determine if a person has heart disease or not. According to World Health Organization, nearly 17.9 million people die due to heart diseases every single year. An Early detection of heart diseases is the only way to avoid and prevent diseases related to heart. A well-known MIT-BIH arrhythmia database is used here. This paper proposes a survey related to automatic analysis and classification of ECG recordings and collectively helps to create a system that can easily and accurately diagnose typical arrhythmias. ECG analysis of larger data is very time consuming, therefore machine learning approaches are proposed for analysis of datasets in this survey. Construction of classification models from a given dataset is easy and their overall performance when we use with different parameters. The proposed approach achieved high accuracy for classifying ECG heartbeats and could be used to assist cardiologists. This survey paper presents a study by analyzing the performance of machine learning algorithms.