DFA Taxonomy for the classification of ECG data for effective health monitoring using ML technology

Abstract

ECG data of patients are collected using sensors which are further classified for monitoring their health. There are certain pitfalls of the existing classification schemes used for health monitoring that are poor extraction of features, ineffective filtering of data, improper access control, and issues related to dimensionality reduction. In this study, Machine learning (ML) is used to perform an early diagnosis of diseases in order to achieve the aim of effective and timely health monitoring of patients. Data preprocessing, Feature extraction, and Activity classification (DFA) are the major components utilised for the implementation of Health monitoring system based on ECG data classification using ML technology. This system classifies recorded activities based on extracted ECG data using Hidden Markov Model (HMM) and Support Vector Machine (SVM) and is integrated with Internet of Medical Things (IoMT) in order to diagnose patient’s disease at early stages. The DFA taxonomy is evaluated based on the effectiveness and performance of the solution. It contributes to the reduction of dimensionalities that facilitates effective feature extraction and improves the accessibility of the model for better health monitoring. The importance of DFA taxonomy is demonstrated by classifying 30 research papers in the domain of health monitoring system. The classification depicts that few components of the ML-based ECG Data Classification system are validated and even fewer are evaluated to depict the effectiveness of the taxonomy.