He, Jinyuan, Rong, Jia, Sun, Le, Wang, Hua, Zhang, Yanchun, and Ma, Jiangang (2018) D-ECG: A dynamic framework for cardiac arrhythmia detection from IoT-based ECGs. In: Lecture Notes in Computer Science (11234) pp. 85-99. From: WISE 2018: 19th International Conference on Web Information Systems Engineering, 12-15 November 2018, Dubai, United Arab Emirates.

PDF (Pubished Version) - Published Version Restricted to Repository staff only

 

14

2

Abstract

Cardiac arrhythmia has been identified as a type of cardiovascular diseases (CVDs) that causes approximately 12 % of all deaths globally. The current progress on arrhythmia detection based on ECG recordings is facing a bottleneck for adopting single classifier and static ensemble methods. Besides, most of the work tend to use a static feature set for characterizing all types of heartbeats, which may limit the classification performance. To fill in the gap, a novel framework called D-ECG is proposed to introduce dynamic ensemble selection (DES) technique to provide accurate detection of cardiac arrhythmia. In addition, the proposed D-ECG develops a result regulator that use different features to refine the classification result from the DES technique. The results reported in this paper have shown visible improvement on the overall heartbeat classification accuracy as well as the sensitivity of disease heartbeats.

Actions (Repository Staff Only)

Item Control Page