Related Experiment Videos
A short-time multifractal approach for arrhythmia detection based on fuzzy neural network
1Department of Biomedical Engineering, Shanghai Jiao Tong University, China. ywang76@sh163c.sta.net.cn
IEEE Transactions on Bio-Medical Engineering
|September 6, 2001
Summary
We developed a new method using short-time multifractality for arrhythmia detection. This approach accurately identifies cardiac arrhythmias like atrial fibrillation and ventricular fibrillation with over 97% accuracy.
Area of Science:
- Cardiology
- Biomedical Engineering
- Data Science
Background:
- Cardiac arrhythmias pose significant diagnostic challenges.
- Existing arrhythmia detection methods may lack accuracy or speed.
- Multifractal analysis offers a novel perspective on complex time series data.
Purpose of the Study:
- To introduce a novel arrhythmia detection approach based on short-time multifractality.
- To enhance classification accuracy using a new fuzzy Kohonen network.
- To evaluate the clinical utility and real-time detection capabilities of the proposed method.
Main Methods:
- Characterization of cardiac rhythms using short-time generalized dimensions (STGDs).
- Discrimination of arrhythmias via a neural network, specifically a novel fuzzy Kohonen network.
- Validation using 180 electrocardiogram records (60 each of atrial fibrillation, ventricular fibrillation, and ventricular tachycardia).
Main Results:
- The proposed algorithm achieved high accuracy, exceeding 97% in classifying different types of arrhythmias.
- The method demonstrated computational efficiency, enabling fast detection.
- The fuzzy Kohonen network improved upon classical algorithms for classification accuracy.
Conclusions:
- Short-time multifractality provides a robust framework for cardiac rhythm analysis.
- The developed algorithm offers a promising, accurate, and fast solution for clinical arrhythmia detection.
- The novel fuzzy Kohonen network enhances the performance of arrhythmia classification systems.