Research Of ECG Signal Classification Algorithm Based On Hybrid Neural Network And Software Design

With the development of society and economy,the national lifestyle has undergone profound changes.Especially with the aging of the population and the acceleration of urbanization,the prevalence of cardiovascular risk factors in China is obvious,which has led to a continuous increase in the number of cardiovascular diseases.The electrocardiogram is an important method for modern medical detection and diagnosis of cardiovascular diseases.It can reliably reflect the activity state of the heart.Traditional electrocardiogram detection requires professional doctors' strong basic knowledge and a lot of clinical experience,and it is precisely because of this method.This process is cumbersome,dependent on labor,and uneven distribution of medical resources.Therefore,it is particularly important to construct an ECG signalassisted diagnosis method with high accuracy and sensitivity.Aiming at the problem that the overall accuracy of the ECG signal classification process is high,but the sensitivity of some types is low.This paper uses the deep learning method to establish a parallel hybrid neural network model based on CNNBi LSTM.On the basis of the convolutional neural network,a bidirectional long and short-term memory neural network is added to enhance the ability of extracting time series features and effectively integrate the heart The spatial and temporal characteristics of electrical signals make up for the lack of ECG signal feature extraction by convolutional neural networks,and realize the automatic classification of ECG signals.By introducing the attention mechanism,the feature extraction and screening capabilities of the hybrid neural network model are further strengthened.Using the MIT-BIH(Massachusetts Institute of Technology Arrhythmia Database)for testing,the classification accuracy of the hybrid neural network with the attention mechanism is 99.31%,and the sensitivity of the five classifications of N,S,V,F,and Q is 99.56,97.21%,98.32%,98.24%,98.19%.Compared with the traditional convolutional neural network,the accuracy rate has increased by 0.69%,and the sensitivity of the five categories of N,S,V,F,and Q has increased by 0.43%,5.89%,3.96%,10.95%,and 3.81% respectively.According to the analysis of actual needs,the ECG signal classification software is designed on the basis of the improved hybrid neural network model.It is connected with wearable devices through Bluetooth to realize the auxiliary diagnosis of cardiovascular diseases in various environments.The disease plays a very good preventive role,reduces the probability of severe cardiovascular disease,and has very important practical significance.