[A study on clustering analysis of arrhythmias]

Area of Science:

• Cardiology
• Biomedical Engineering
• Data Science

Context:

• Electrocardiogram (ECG) analysis presents challenges due to large data volumes, high accuracy requirements, and the need for real-time processing.
• Existing arrhythmia detection methods may not fully capture individual variations or integrate waveform and rhythm analysis effectively.

Purpose:

• To develop and validate a novel, classified algorithm for arrhythmia detection based on clustering analysis.
• To leverage the "things-of-one-kind-come-together" principle for grouping similar cardiac conditions.
• To incorporate individual differences into arrhythmia analysis through QRS complex waveform and rhythm analysis.

Summary:

• The proposed algorithm classifies arrhythmias by clustering ECG data, focusing on QRS complex waveforms and rhythm analysis.
• It accounts for individual patient variations, enhancing the precision of arrhythmia identification.
• Validation using eight records from the MIT-BIR standard heart electricity database demonstrated a correct clustering probability exceeding 90%.

Impact:

• Provides an effective and accurate method for real-time arrhythmia analysis.
• Enhances diagnostic capabilities in cardiology through advanced data clustering techniques.
• Offers a foundation for developing more sophisticated automated ECG interpretation systems.