TECHNOLOGY FOR HEALTH BENEFITS: CLASSIFICATION OF ELECTROCARDIOGRAM SIGNALS USING ARTIFICIAL INTELLIGENCE

Abstract

Ventricular extrasystoles are a type of arrhythmia characterized by isolated ventricular impulses and can affect both healthy individuals and those with heart conditions. Hence, the importance of diagnosis, which is typically made by a medical team through an electrocardiogram (ECG) examination of the patient. Considering this, the use of two artificial intelligence methods for classifying ECG signals is proposed to assist these professionals: Multiple Linear Regression (MLR) and k-means algorithm. Besides that, MIT-BIH Arrhythmia Database was used, comprising a total of 155 heartbeats from a patient, and an accuracy of 74.19% was achieved using both MLR and k-means. Thus, it has been demonstrated that the application of MLR and k-means in the classification of heartbeats from the ECG examination, combined with the evaluation of the medical team, can lead to improved analyses in the diagnosis of ventricular extrasystoles.