ECG classification efficient modeling with artificial bee colony optimization data augmentation and attention...

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Healthcare
• Signal Processing

Background:

• Electrocardiogram (ECG) signal analysis faces challenges due to data imbalance and suboptimal modeling techniques.
• Existing methods often struggle with accurate identification, classification, and diagnosis of cardiac conditions from ECG data.

Purpose of the Study:

• To improve the performance and accuracy of ECG signal modeling.
• To address data imbalance issues in ECG datasets.
• To optimize hyperparameter configurations for enhanced ECG classification.

Main Methods:

• Employed TimeGAN (Generative Adversarial Network) for data augmentation and balancing of ECG signals.
• Utilized an artificial bee colony optimization algorithm to fine-tune model hyperparameters by minimizing Wasserstein distance.
• Integrated an Efficient network with attention mechanisms to enhance modeling performance.

Main Results:

• Data augmentation using TimeGAN significantly improved classification accuracy to 99.51%.
• The optimized model achieved an overall accuracy of 99.70% and an average positive prediction rate of 99.44%.
• The integrated approach effectively addressed challenges in ECG signal identification, classification, and diagnosis.

Conclusions:

• The combination of TimeGAN and artificial bee colony optimization with Efficient networks offers a robust solution for ECG signal analysis.
• This methodology significantly enhances diagnostic accuracy and overcomes limitations of imbalanced datasets in cardiac monitoring.
• The developed approach shows great promise for improving automated ECG interpretation in clinical settings.