ResNet-50-CNN and LSTM Based Arrhythmia Detection Model Based on ECG Dataset

The ECG is a critical component of computer-aided arrhythmia detection systems since it helps to reduce the rise in the death rate from disorders of the circulatory system. However, due to the intricate changes and imbalance of electrocardiogram beats, this is a difficult problem to solve. This study provides an innovative and enhanced ResNet-50 model using a Conv-1D model with Long Short Term Memory (LSTM) based on Convolution Neural Network (CNN) approach for arrhythmia identification using ECG data, including proper parameter optimization and model training. The results of applying the proposed model to the MIT-BIH arrhythmia database demonstrates that the model performs better, having an accuracy of 98.7% and a MSE of 0.06 when compared to other classification methods.