ECG Signal Classification Using Recurrence Plot-Based Approach and Deep Learning for Arrhythmia Prediction

Automatic electrocardiogram (ECG) analysis is crucial in diagnosing heart arrhythmia but is limited by the performance of existing models owing to the high complexity of time series data analysis. Arrhythmia is a heart condition in which the rate or rhythm of the heartbeat is abnormal. The heartbeat may be excessively fast or slow or may have an irregular pattern. Research has shown that the use of deep Convolutional Neural Networks (CNNs) for time-series classification has several advantages over other methods.They are highly noise-resistant models and can very informatively extract deep features that are independent of time. Five classes of heartbeat types in the MIT-BIH arrhythmia database were classified using the resilient and efficient deep CNNs proposed in this study. The proposed method achieved the best score (95.8% accuracy) for arrhythmia detection using the deep learning classification method.