ECG Classification Using Combination of Linear and Non-Linear Features with Neural Network.

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• Electrocardiogram (ECG) analysis is crucial for diagnosing cardiac conditions.
• Current ECG classification methods face challenges in accuracy and reliability.
• Distinguishing between normal and abnormal heartbeats requires sophisticated feature analysis.

Purpose of the Study:

• To develop an advanced ECG classification approach combining linear and non-linear dynamics.
• To improve the accuracy and reliability of automated ECG abnormality detection.
• To establish a novel method for feature extraction from ECG signals.

Main Methods:

• Feature analysis integrating linear and non-linear ECG dynamics.
• Utilizing complexity measures and ordinal network analysis for non-stationarity assessment.
• Applying ECG partitioning techniques, specifically on PQRST complex data.

Main Results:

• Demonstrated effective detection of cardiac abnormalities through automatic feature extraction.
• Achieved improved state-of-the-art performance on the standard ECG5000 dataset.
• Validated the proposed technique's capability in enhancing ECG classification accuracy.

Conclusions:

• The combined analysis of linear and non-linear ECG dynamics offers a robust approach for improved classification.
• The method shows significant potential for clinical application in automated cardiac diagnostics.
• Further research can explore broader datasets and diverse cardiac conditions.