Comparison of real-time classification systems for arrhythmia detection on Android-based mobile devices

Area of Science:

• Cardiology
• Biomedical Engineering
• Machine Learning

Background:

• Electrocardiogram (ECG) is crucial for diagnosing heart conditions like ischemia and arrhythmias.
• Mobile devices offer a platform for accessible, low-cost, real-time health monitoring.
• Developing automated systems for ECG analysis on smartphones is a key area of research.

Observation:

• The study analyzed MIT-BIH Arrhythmia and Supraventricular Arrhythmia databases.
• R-peaks were detected using a QRS algorithm, and 16 features were extracted.
• Eight classifiers were trained and evaluated for computational cost and memory demand.

Findings:

• The C4.5 classifier demonstrated the highest accuracy (91.6%) for distinguishing normal and abnormal heartbeats.
• This classifier requires further feature selection evaluation for optimization.
• The system is designed for real-time classification on Android mobile devices.

Implications:

• This research paves the way for widespread, affordable, mobile ECG monitoring.
• Early detection of arrhythmias via smartphones can improve patient outcomes and reduce emergency department visits.
• Further development could integrate this system into daily health management tools.