Machine learning tools for Diagnosis of Heart Arrhythmia

Title: Machine Learning Tools for Diagnosis of Heart Arrhythmia Author: Glejdis Shkëmbi Department / Institute: Department of Theoretical Computer Science and Mathematical Logic Supervisor of the bachelor thesis: Mgr. Marta Vomlelová, Ph.D., Department of Theoretical Computer Science and Mathematical Logic Abstract: Electrocardiogram (ECG) is considered to be the most reliable, efficient and low-cost tool used in the healthcare industry to diagnose cardiac arrhythmia. However, visual representation of ECG signals manually by medical workers is intricate and time-consuming, and may lead to human mistakes and inaccuracy in heartbeat recognition. In this paper, different machine learning techniques for the classification of five classes of ECG heartbeats using Discrete Wavelet Transform (DWT) features are compared. In particular, the significant role of statistical features of DWT coefficients in distinguishing between different heartbeat classes is highlighted. Performances of the models have been evaluated using the online MIT-BIH arrhythmia database. The obtained results indicate the reliability of the machine learning-based approaches for diagnoses of cardiac arrhythmia from ECG signals. Keywords: Electrocardiogram (ECG); Discrete Wavelet Transform (DWT); Support Vector Machine (SVM); Random Forest; Heart..