Visualization Analysis Of Knowledge Map Of Chronic Bronchitis Based On Deep Learning

Chronic bronchitis,also known as chronic bronchitis,is a chronic non-specific inflammation that often leads to serious symptoms such as emphysema and chronic pulmonary heart disease,posing a serious threat to the health of patients.The treatment of chronic bronchitis with traditional Chinese medicine has advantages such as overall syndrome differentiation and treatment of both symptoms and symptoms.The records of the diagnosis and treatment of chronic bronchitis by renowned doctors contain the thinking methods and academic ideas of traditional Chinese medicine.However,traditional Chinese medicine medical records often use text as the carrier,and TCM terms such as symptoms and syndrome types may have different n ames in different medical records.The lack of standardization and expression leads to a lack of unified standards when annotating texts.These personalized medical records increase the difficulty of data processing.With the application of artificial inte lligence technology,the construction of knowledge map is widely used in the medical field.Knowledge map can discover the tacit knowledge in the medical records of traditional Chinese medicine,and reveal the treatment and medication rules for specific di seases and specific syndromes.This study is based on the method of knowledge graph to standardize the entities in traditional Chinese medicine medical records.Through experimental model comparison,the BERT-BILSTM-CRF model is selected for improvement,and a joint extraction model is constructed to extract entity relationships from medical records.The extracted data is then visualized through the Neo4 j graph database,and the knowledge graph visualization analysis is conducted from two levels of disease and syndrome types.The main research content of this article is as follows:(1)Standardize the medical terms in the text through traditional Chinese medicine terminology standards,so as to unify text entities from different sources.Four named entity recognition models,BERT,BILSTM,BILSTM CRF,and BERT BILSTM CRF,were compared,and the superiority of the BERT BILSTM CRF model was verified through experimental results.This model first uses the BERT model as a pre trained Chinese language model to effi ciently extract rich semantic features contained in the context of traditional Chinese medicine medical records,and then inputs feature vectors into the BILSTM-CRF model to extract deep level text features and achieve entity recognition.(2)In the part o f entity relationship extraction,based on the named entity recognition model and through model improvement,this paper constructs a joint entity relationship extraction model based on parameter sharing to extract knowledge extraction from medical record t ext data.The construction principle of the model is to take the output of BILSTM model as the shared coding layer,realize the sharing of underlying parameters,and take the splicing of two subtasks loss function as the final loss function,so as to impro ve the performance of relationship extraction.Finally,the entity relationship triplet information in the medical record is extracted through this model.(3)In the visualization section of the knowledge graph of chronic bronchitis,based on the character istics of traditional Chinese medicine medical records,the Neo4 j graph database is selected to visualize the extracted entity triplet information.Through the Neo4 j graph database,the knowledge graph "disease symptom tongue pulse syndrome type formula tr eatment principle treatment" is visualized,and the correlation between medical record entities is intuitively felt through visual graphics.After visualizing the knowledge graph,guided by the syndrome differentiation and treatment system of traditional C hinese medicine,entities and relationships in medical records are analyzed from the disease level and syndrome type level through graph database query language,in order to better understand and predict the pathogenesis of diseases.Based on the above res earch content,this article designs and implements the construction of a knowledge map of traditional Chinese medicine medical records for chronic bronchitis.Through a deep learning model,the processing of medical record text data was completed,and data conversion and entity relationship extraction were achieved.The joint presentation of tacit knowledge and explicit knowledge is realized through Neo4 j diagram database.The construction of this knowledge graph has opened up a new direction for better inh eriting traditional Chinese medicine,assisting physicians in more accurate diagnosis and treatment,bringing more benefits to society,and promoting the intelligent process of traditional Chinese medicine.