Title: Maximisation of arrhythmia classification accuracy by addressing class overlap and imbalance problem
Authors: R. Rekha; R. Vidhyapriya
Addresses: Department of Information Technology, PSG College of Technology, Coimbatore 641004, India ' Department of Information Technology, PSG College of Technology, Coimbatore 641004, India
Abstract: Automation in arrhythmia classification helps medical professionals to make accurate decisions upon the patient's health. Classification becomes complicated when class overlapping and class imbalance problem occurs together. The aim of this work is to improve the arrhythmia classification accuracy. Proposed methodology consists of fisher discriminant ratio based feature ranking stage and anomaly detection based training sample selection stage followed by classification using probabilistic neural network classifier. As per the recommendations of the Association for the Advancement of Medical Instrumentation, five arrhythmia classes were classified. The proposed method resulted in average sensitivity, positive predictive value and F Score of 95.37%, 98.35% and 96.72%, respectively. The experimental results revealed that: (1) Selected non-overlapping features were able to better discriminate arrhythmia classes, (2) Mixture of Gaussians based anomaly detection method suited well to handle the class imbalance problem and (3) Minority classes with few training samples were also correctly classified using the proposed method.
Keywords: arrhythmia classification; feature selection; class overlap; class imbalance; probabilistic neural network classifier; AAMI; MIT-BIH arrhythmia database.
International Journal of Biomedical Engineering and Technology, 2018 Vol.26 No.2, pp.197 - 216
Available online: 06 Jan 2018 *Full-text access for editors Access for subscribers Purchase this article Comment on this article