Monitoring and alerting system for heart abnormality by ECG

Abstract

For cardiovascular disease prediction, a variety of Machine Learning (ML) algorithms are increasingly being utilized. “The predictive ability of ML algorithms in cardiovascular diseases is promising, particularly Support Vector Machine (SVM) and boosting algorithms”. However, since the heart problem is complicated, and the equipment requirements for more complex heart diseases are correspondingly higher, this project aims to provide a portable monitoring and prediction services for milder heart diseases. The key components of this system is a hardware-based biosensor with algorithms, that are targeted to detect anomalies and predict the probability of the patient having arrhythmia and coronary artery disease. Waveform segmentation algorithms are used to better process the benchmark dataset and normal heartbeat, followed by data pre-processing, and lastly datasets are used to train the models. This report will discuss the entire process of completing this monitoring and alerting system, from motivations, system structure, hardware selection and setup to software development.