Research Of Multi-label Classification Algorithm For ECG-assisted Diagnosis

Since the 21st century,heart disease and hypertension have threatened human health.They are two diseases with extremely high morbidity and mortality.At the same time,artificial intelligence technologies such as machine learning and deep learning are also developing rapidly in the field of smart medicine.The use of electrocardiogram(ECG)data for the diagnosis of heart diseases has achieved certain results,but there are problems in clinical applications such as incomplete diagnosis categories and low accuracy of some categories.This paper proposes a multi-label ECG assisted diagnosis algorithm based on deep learning.On the one hand,it solves the problem of category imbalance that is common in multi-label scenarios,and on the other hand,it improves the comprehensiveness and accuracy of diagnostic categories.And the algorithm is applied to the early diagnosis of hypertensive heart disease and the ECG intelligent assisted diagnosis system.The main work and research content are divided into the following parts:1.Aiming at the ECG data set,a multi-label heart disease assisted diagnosis algorithm based on deep learning is designed.The algorithm combines the residual neural network and the gated recurrent unit network model to make full use of the timing and local features of the ECG sequence.At the same time,aiming at the problem of multi-dimensional category imbalance in multi-label scenarios,improved Focal Loss was proposed as the Loss function,and the multi-label mutual exclusion relationship was combined with the objective function to accelerate the model training speed.The proposed algorithm has a significant improvement in multiple indicators such as F1-Score,AUC value and Hamming loss.2.Due to the clinical correlation between heart disease and hypertension,this paper applies the multi-label ECG assisted diagnosis algorithm based on deep learning to the hypertension and ECG data set,and designs a correlation mining method based on the decision tree algorithm.The relationship between hypertension and the clinical manifestations of a variety of heart diseases is obtained,which provides a basis for the early diagnosis of hypertensive heart disease.3.This paper completes the construction of the ECG intelligent assisted diagnosis system,and designs the system in detail from functional requirements,front-end and back-end,model deployment,and implements various business modules.The system uses Spring boot and Mybatis framework,MySQL database.This system uses the Flask framework to deploy the heart disease assisted diagnosis model for the Web system to call.The front-end display system is constructed through the Vue framework.