Related Experiment Video
Updated: Jul 3, 2026
Selecting Multiple Biomarker Subsets with Similarly Effective Binary Classification Performances
Published on: October 11, 2018
Multi-expert ensemble ECG diagnostic algorithm using mutually exclusive-symbiotic correlation between 254
Jiewei Lai1,2, Yue Zhang1,2, Chenyu Zhao1,2
1School of Biomedical Engineering, Southern Medical University, Guangzhou, China.
Insights
A new multi-expert ensemble learning model can identify 254 electrocardiogram (ECG) terms, significantly improving heart health diagnostics. This advanced AI tool offers comprehensive support for public health by detecting more arrhythmias than previous methods.
Area of Science:
- Cardiology
- Artificial Intelligence
- Machine Learning
Background:
- Electrocardiograms (ECGs) are crucial for heart health assessment but current AI models detect limited conditions.
- Existing intelligent ECG diagnostic tools cover only a few common arrhythmias, necessitating further clinical review.
Purpose of the Study:
- To develop an advanced multi-expert ensemble learning model for comprehensive ECG analysis.
- To enhance the diagnostic capabilities of AI in recognizing a wider spectrum of ECG abnormalities.
Main Methods:
- Development of a multi-expert ensemble learning model trained on 191,804 wearable 12-lead ECGs.
- Application of mutually exclusive-symbiotic correlations between hierarchical multiple labels at the loss level.
- Addressing class imbalance challenges to improve model robustness.
Main Results:
- The model successfully recognizes 254 distinct ECG terms.
- Achieved high performance with an average area under the receiver operating characteristic curve of 0.973 (offline) and 0.956 (online).
- Selected 130 clinically significant terms for practical application.
Conclusions:
- The developed model offers real-time, comprehensive ancillary support for public ECG interpretation.
- This AI-driven approach significantly advances the diagnostic accuracy and scope for ECG analysis.
- The model provides valuable support for cardiologists and public health initiatives.
Abstract:
Electrocardiograms (ECGs) are a cheap and convenient means of assessing heart health and provide an important basis for diagnosis and treatment by cardiologists. However, existing intelligent ECG diagnostic approaches can only detect up to several tens of ECG terms, which barely cover the most common arrhythmias. Thus, further diagnosis is required by cardiologists in clinical settings. This paper describes the development of a multi-expert ensemble learning model that can recognize 254 ECG terms. Based on data from 191,804 wearable 12-lead ECGs, mutually exclusive-symbiotic correlations between hierarchical multiple labels are applied at the loss level to improve the diagnostic performance of the model and make its predictions more reasonable while alleviating the difficulty of class imbalance. The model achieves an average area under the receiver operating characteristics curve of 0.973 and 0.956 on offline and online test sets, respectively. We select 130 terms from the 254 available for clinical settings by considering the classification performance and clinical significance, providing real-time and comprehensive ancillary support for the public.
