A High-Performance Anti-Noise Algorithm for Arrhythmia Recognition
Robustness
Statistic
Feature (linguistics)
Cardiac arrhythmia
DOI:
10.3390/s24144558
Publication Date:
2024-07-15T18:15:49Z
AUTHORS (5)
ABSTRACT
In recent years, the incidence of cardiac arrhythmias has been on rise because changes in lifestyle and aging population. Electrocardiograms (ECGs) are widely used for automated diagnosis arrhythmias. However, existing models possess poor noise robustness complex structures, limiting their effectiveness. To solve these problems, this paper proposes an arrhythmia recognition system with excellent anti-noise performance: a convolutionally optimized broad learning (COBLS). proposed COBLS method, signal is convolved blind source separation using analysis method based high-order-statistic independent component (ICA). The constructed feature matrix further feature-extracted dimensionally reduced principal (PCA), which reveals essence signal. linear correlation between data can be effectively reduced, redundant attributes eliminated to obtain low-dimensional that retains essential features classification model. Then, realized by combining (BLS). Subsequently, model was evaluated MIT-BIH database stress test database. outcomes experiments demonstrate exceptional performance, impressive achievements terms overall accuracy, precision, sensitivity, F1-score. Specifically, results indicate outstanding figures reaching 99.11% 96.95% 89.71% 93.01% F1-score across all four experiments. shows 24 dB, 18 12 dB signal-to-noise ratios.
SUPPLEMENTAL MATERIAL
Coming soon ....
REFERENCES (38)
CITATIONS (1)
EXTERNAL LINKS
PlumX Metrics
RECOMMENDATIONS
FAIR ASSESSMENT
Coming soon ....
JUPYTER LAB
Coming soon ....