Comparison of different classification methods in arrhythmia detection using Hjorth descriptors

摘要

Telemedicine starts to be beneficial to patients in remote regions. Monitoring the ECG signals of these patients is very important. Developments in information technology start to provide important contribution to the clinical decision support systems for early detection and diagnosis. This study aimed to identify ECG arrhythmia by using Hjorth descriptors as main features. Different classification methods are compared using these features. For classification, Matlab Classification Learner and Neural Network Toolbox are used. Cubic SVM, Quadratic SVM, Complex Tree, Fine KNN, Boosted Trees, Subspace Discriminant, Neural Network are used as classification methods. The performance of the method is tested on data used obtained from the PhysioNet database.