Abnormality Heartbeat Classification of ECG Signal Using Deep Neural Network and Autoencoder

Electrocardiogram (ECG) is a device used by healthcare practitioners to monitor and processing of patient health data so can detect abnormality cardiovascular disease. Continuous heart supervision generates large amounts of data and analyzes this large data need classification method. This Paper exposes the classification of heartbeat abnormality based on the ECG signal by using Deep Neural Network (DNN). Three preprocessing stages of the ECG signal are applied before the classification process, which is segmentation, normalizing using normalize bound, and feature extraction by using Autoencoder. The results show that the applied method gets an outstanding accuracy about 99.22% and sensitivity about 98.03%.