Design And Implementation Of Anauxiliary Diagnosis And Treatment System Based On Electronic Medical Records In Neurology Department

In recent years,the morbidity and mortality of neurological diseases have been on the rise,and affected by the development of different regions,the distribution of medical resources is seriously uneven.Therefore,paramedical technology for medical staff becomes particularly important.This paper studies the related theories and algorithms of auxiliary diagnosis and treatment,uses the method of intelligent disease derivation and similar medical record retrieval to design the auxiliary diagnosis and treatment system of this paper,through the doctor input medical record text to deduce the possible disease,and can retrieve the similar medical record in the medical record database,so as to assist in diagnosis.The main research contents of this paper are as follows:(1)Research and application of named entity recognition based on Bert IDCNN-CRF.This model only trains IDCNN-CRF layer with Bert parameters unchanged,which reduces the training parameters and improves the speed of training while solving the ambiguity of words.(2)Research and application of disease derivation model based on named entity recognition results and Bayesian network in Neurology department.Different from general statistical methods,the most important point of Bayesian probability statistical model is that it can make full use of the prior information of data and has excellent accuracy and interpretability.In this paper,bayesian probabilistic inference and deep learning are combined to construct a local Bayesian network structure for inference based on named entity recognition results,which simplifies inference calculation but ensures the accuracy of inference.(3)Based on the research and application of similar medical records with named entity recognition results and disease derivation results in neurology department,the graph embedding node2Vec algorithm was adopted to overcome the problem of sparse information contained in medical records text.Through the graph embedding method on the network and the electronic medical record data mapping vectors study,through setting the random walk of strategy unearth hidden key relationship between entities in the network,then access to electronic medical records of vector,said by cosine similarity computing text similarity,similarity is obtained by dading heap sort again the list of similar records from high to low.(4)Design and implement the auxiliary diagnosis and treatment system based on the electronic medical record of neurology department,which integrates medical record management,disease derivation and similar medical record retrieval,deduces possible diseases through the correlation between medical entities and diseases contained in the medical record,calculates the similarity between medical records,and realizes the retrieval of similar medical records.And to modify,store,use,transmission management of medical records.