Design And Implementation On Early Detection System Of Myocardial Ischemia Via Cardiodynamicsgram

Myocardial ischemia is a common cardiovascular disease that has become one of the most important factors affecting the health of the nation.Studies have shown that timely detection and treatment of myocardial ischemia is the key to reducing premature mortality from cardiovascular disease.The diagnosis of myocardial ischemia is mainly based on the relevant clinical information in the electronic medical record and electrocardiogram(ECG).The data in the electronic medical record is complex and the keywords are vague,so it takes a lot of effort to extract the test results from it manually.At the same time,because of the small amplitude of the ECG signal and insignificant characteristics,the diagnostic accuracy of doctors interpreting ECG for myocardial ischemia is not high.To address these problems,this paper proposes a medical entity recognition method to extract key information such as examination results in electronic medical records and model the dynamics of ECG signals based on deterministic learning to obtain cardiodynamicsgram(CDG),which in turn enables early detection of myocardial ischemia.The main work is as follows.1.In order to extract examination information related to myocardial ischemia from electronic medical records containing a large amount of text,this paper proposes a named entity recognition method based on adaptive word embedding.First,an active learning method combining uncertainty sampling and text similarity is used to filter out valuable texts from unlabeled medical data for labeling,which in turn leads to a self-labeled entity recognition dataset.Then,three sets of word vectors,Bigram,BERT hidden vector and Soft-lexicon,are generated from the input text and input to the Bi LSTM+CRF model for further training.Finally,the effectiveness of the named entity recognition method based on adaptive word vectors is verified on the self-labeled dataset and the public dataset.2.In order to solve the problem of low accuracy of ECG in early detection of myocardial ischemia,this paper proposes an early detection method of myocardial ischemia based on CDG.First,the myocardial ischemia dataset is constructed by combining medical entity recognition methods.Then,deterministic learning is used to model the dynamics of ECG signals and generate CDG.Then,a combination of nonlinear dynamics analysis and convolutional neural network is used to extract features from the CDG and construct a myocardial ischemia detection model.Finally,the method was validated on the self-built and publicly available datasets with an accuracy of 88.60% and 96.00%,respectively.3.In order to assist medical personnel in data management and detection of myocardial ischemia,an early detection system of myocardial ischemia is built in this paper.The system can support case information management,ECG data management,and data analysis and diagnosis,and realize the automatic parsing and key information extraction of electronic medical records,automated processing and modeling of ECG data,and calculation of myocardial ischemic indexes.