Enhanced Heart Arrhythmia Detection Using Ensemble Learning and Statistical Wave Interval Analysis: A Proposal for Mobile Integrated Diagnostic Systems

In today’s modern era, heart attack and heart arrhythmia stand out as two prevalent and significant cardiac concerns that can impact individuals of any age. They result in harm to the heart’s muscle tissue by obstructing blood flow. In the case of heart arrhythmia, irregular heartbeats can compromise the heart’s ability to effectively pump blood throughout the body. According to the 2023 World Health Report, there were 2.1 million global deaths attributed to heart arrhythmia in 2022, rising to 2.2 million in 2023, reflecting a 4.8% increase globally. Diagnosis typically involves an electrocardiogram (ECG), which measures the heart’s electrical activity and can detect various heart conditions such as arrhythmia, heart attacks, or heart failure. Our study focuses on categorizing five types of heartbeats using a dataset obtained from MIT-BIH through the application of ensemble learning techniques, particularly the Bagging Classifier algorithm. Our results demonstrate a remarkable accuracy rate of 99.67% in arrhythmia classification. By combining ensemble learning with feature normalization methods, we have achieved significant improvements in classification accuracy, leading to enhanced prediction of arrhythmia, proactive management of cardiac issues, and ultimately, better patient outcomes.