Hybrid classification of Bayesian and extreme learning machine for heartbeat classification of arrhythmia detection
- ,
- Naomie Salim,
- Taqwa Ahmed,
- Arief R. Harris,
- Tan Tian Swee
- Universiti Teknologi Malaysia
Abstract
A cardiac circumstance affected through irregular electrical action of the heart is called an arrhythmia. A noninvasive method called Electrocardiogram (ECG) is used to diagnosis arrhythmias or irregularities of the heart. The difficulty encountered by doctors in the analysis of heartbeat irregularities id due to the non-stationary of ECG signal, the existence of noise and the abnormality of the heartbeat. The computer-assisted study of ECG signal supports doctors to diagnoses diseases of cardiovascular. The major limitations of all the ECG signal analysis of arrhythmia detection are because to the non-stationary behavior of the ECG signals and unobserved information existent in the ECG signals. In addition, detection based on Extreme learning machine (ELM) has become a common technique in machine learning. However, it easily suffers from overfitting. This paper proposes a hybrid classification technique using Bayesian and Extreme Learning Machine (B-ELM) technique for heartbeat recognition of arrhythmia detection AD. The proposed technique is capable of detecting arrhythmia classes with a maximum accuracy of (98.09%) and less computational time about 2.5s.
Research Output Information
Output type
Host publication Subtitle
Elevating Community Through ICT, ICT-ISPC 2017Original language
EnglishPages from-to (Number of pages)
Pages 1-4 (4 pages)Publication milestones
- Published - 19/10/2017
Publication status
Publisher
Institute of Electrical and Electronics Engineers Inc.Publication series
- Publication series name: 6th ICT International Student Project Conference: Elevating Community Through ICT, ICT-ISPC 2017
Volume: 2017-January
ISBN (Electronic)
9781538629963Publication IDs
- Scopus: 85043690170
