Heart Diseases Classification Using Machine Learning Techniques

Abstract

Abstract: Heart is the most vital organ in the human body, generating a systematic time-varying signal due to its electrical activity is called as an electrocardiogram (ECG). An electrocardiogram records the electrical signals in the heart. It's a common and painless test used to quickly detect heart problems and monitor the heart's health. It is a well-established diagnostic tool for cardiac diseases. ECG signal is monitored by placing sensors as positions on chest and limb. Each heart beat is caused by a section of the heart generating an electrical signal. Nowadays, heart diseases classification is one of the vital problems in health care sector. Therefore, this work aims to classify different heart diseases using machine learning techniques (such as linear discriminant analysis, support vector machine (SVM), multilayer perceptron, random forest, k-nearest neighbour). In order to classify the signal, it consists of major two steps: first step is to pre-process or extract the features and second stage is to apply machine learning algorithms. The performance of these methods can be assessed by popular MIT-BIH Arrhythmia database and quantitative metrics such as accuracy, sensitivity, precision.