Hybrid Deep Learning Approach For Ecg Signal Classification Using Sd-Rom

Abstract

Electrocardiogram (ECG) data show how the heart beats electrically and are used to diagnose and monitor heart problems. Advanced algorithms and deep learning methods can make ECG data analysis much more accurate, leading to better patient results. In this work, a new way to group ECG data is shown. In the proposed system, the Signal Dependent Rank Order Mean (SD-ROM) method, Recurrent Neural Networks (RNNs), and Convolutional Neural Networks (CNNs) all work together. The SD-ROM method cleans up and improves the ECG data by highlighting essential parts and lowering noise. After being worked on, the signs are sent to a CNN-RNN model. The CNN part looks for marks in space, and the RNN part looks for links between events over time. This method takes the best parts of both CNN and RNN systems and puts them together to get very good classification accuracy. Experiments show that the SD-ROM-enhanced CNN-RNN model is better at classifying things than other methods. It provides a solid way to look at ECG data in real-time and helps find early heart diseases.