Research On ECG Intelligent Diagnosis Technology Based On Multi-label Classification

With the development of technology,the technology of intelligent analysis of ECG came into being.Due to the complexity of electrocardiogram recognition,the computer automatic analysis and diagnosis system has not yet reached the acceptable level.The difficulty of self-intelligent diagnosis of ECG mainly lies in the following two aspects:(1)The ECG signal is complex and easily affected by noise,so it is difficult to effectively extract features from the ECG data.(2)The classification of electrical signals is a multi-label classification problem.Considering the diversity of labels,it is difficult to accurately classify the ECG signal.In order to solve the above problems,the main research work of the paper are as follows.(1)In view of the fact that P wave has low amplitude,which makes it easy to be interfered by noise.Thus,it is difficult to accurately detect the P wave.This paper proposes a multi-lead joint P wave detection algorithm based on downsampling.Downsampling can greatly reduce the interference of noise on P wave.At the same time,considering that single-lead detection is prone to misdetection,the multi-lead joint detection method is used to correct positioning errors,so as to accurately identify and locate P wave.(2)The classification of ECG signals is a multi-label classification problem.Considering the diversity of label,it is difficult to accurately predict all labels to which a sample belongs.This paper combines multiple different types of classifiers and design voting algorithms to classify ECG signals.In order to make full use of the label correlation,a multi-label classification method based on the optimal similarity of feature space is proposed.This method mines label correlation from label space and feature space.Experimental results show that the method in this paper can classify ECG signals accurately and efficiently.(3)Based on the research of ECG signal processing and multi-label classification algorithm,the ECG multi-label classification prototype system is designed and implemented.The system integrates four modules: data preprocessing,feature extraction,model training and multi-label classification,which can assist doctors in diagnosing patients.