Machine Learning And Ecg-Based Arrhythmia Classification Exploiting R-Peak Detection


  • Van Thinh Pham PTIT


ECG, EEMD, Hilbert transform, Machine learning, Arrhythmia classification


The increasing incidence of heart-related diseases has prompted the development of efficient techniques to identify irregular heart problems. It has proven to be challenging to promptly and accurately diagnose many complicated and interferential symptom diseases including arrhythmia. Thanks to the recent evolution of artificial intelligence (AI) and the advances in signal processing, automated arrhythmia classification has become more effective and widely applied for physicians and practitioners with machine learning (ML) techniques and the use of electrocardiograms (ECG). In this work, we have investigated machine learning-based arrhythmia classification problem based on ECGs and successfully proposed an efficient ECG-based machine learning solution employing R-peaks. In order to enhance the arrhythmia diagnosis performance, our developed approach exploits a Butterworth filter and utilizes EEMD technique, Hilbert transformation and a proper machine learning algorithm. The performance of the proposed method is evaluated with the most popular public dataset, MIT-BIH Arrhythmia. The numerical results imply that the developed method outperforms the notable algorithms given in the conventional works and obtains better performance with the accuracy of 93.4%, the sensitivity of 95.4% and the F1-score of 96.3%. The attained high F1-score proves that the proposed method can effectively deal with the data imbalance while detecting arrhythmia, or in other words, it can be suitable and proper to deploy in practical clinical environments.