A Study on ECG Electrocardiogram Signal Classification Based on Improved CNN-LSTM

According to big data, about 17 million people develop arrhythmias every year, and timely diagnosis can effectively reduce the risk of disease. In this paper, we propose an automatic ECG signal detection algorithm based on a mixture of improved convolutional neural network (CNN) and long short-term memory (LSTM) network, which can classify six different arrhythmic diseases. The ECA attention mechanism is introduced into the CNN-LSTM network, which avoids the dimensionality reduction of the module and improves the ability of the model to capture cross-channel interactive information. At the same time, the Hard Swish function was selected as the activation function of the model in the CNN network, which improved the accuracy of the model, and the experimental results showed that the accuracy of the improved model in the ECG signal classification task was 99.31%, F1 was 98.52%, and the recall rate was 99.01%. The method proposed in this study can efficiently and accurately identify different arrhythmic diseases.