An Arrhythmia Classification Model Based on a CNN-LSTM-SE Algorithm.

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Medicine
• Cardiology

Background:

• Arrhythmia is a primary cause of sudden cardiac death.
• Electrocardiogram (ECG) signal analysis is crucial for noninvasive arrhythmia diagnosis.

Purpose of the Study:

• To develop and evaluate a novel arrhythmia classification model.
• To improve the accuracy and reliability of ECG-based arrhythmia detection.

Main Methods:

• Utilized the MIT-BIH arrhythmia database for training and testing.
• Applied the Ensemble Empirical Mode Decomposition (EEMD) algorithm for ECG signal noise reduction.
• Developed a hybrid model combining Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM) networks, and a Squeeze-and-Excitation (SE) channel attention mechanism.

Main Results:

• The proposed CNN-LSTM-SE model demonstrated superior performance compared to LSTM, CNN-LSTM, and LSTM-attention models.
• Achieved a high classification accuracy of 98.5%.
• Exhibited excellent performance metrics, including precision (>97%), recall (>98%), and F1-score (>0.98) across all labels.

Conclusions:

• The CNN-LSTM-SE model effectively classifies arrhythmias from ECG data.
• The model's high accuracy and performance metrics indicate significant practical value for clinical arrhythmia prediction.