Construction Of Ethnomedicine Knowledge Graph Based On BERT-BiLSTM-CRF Knowledge Extraction Model

Chinese Ethnic medicine is an indispensable part of traditional Chinese medicine,and its diagnosis and treatment are gradually integrated into mainstream Western medicine.However,there are few reports related research and intelligent applications of knowledge graphs in the field of Chinese ethnic medicine.At present,Chinese ethnic medicine has accumulated a lot of research results in the prevention and treatment of major diseases,but there is no data set available in the open Internet.“DICTIONARY OF CHINESE ETHNIC MEDICINE” is an authoritative monograph in the field of Chinese ethnic medicine;it has carried out a comprehensive analysis and collation based on the existing ethnic medicine research literature,while the text information organized by this dictionary has not been formatted and formed a database that is easy to query.If the knowledge graph is used to store,organize,and manage data such as Chinese ethnic medicine in this dictionary,it will definitely promote the intelligent application of knowledge in the field of Chinese ethnic medicine.Extracting these unstructured texts will inevitably bring a lot of time and financial consumption.To this end,this paper proposes a knowledge extraction model to automatically and efficiently extract the knowledge in this dictionary,and provide a comprehensive and reliable data source for the construction of the knowledge graph of Chinese ethnic medicine.In view of this,this article carries out the following research work:1.Based on BERT-Bi LSTM-CRF and the annotation and review data of relevant professionals,a knowledge extraction model is proposed that extracts relationships first and then entities.The output of the first relationship extraction model is the input of the second entity extraction model.They are combined into a BERT-Bi LSTM-CRF knowledge extraction model;through the comparison of model performance,it is shown that the model can efficiently extract the relations and entities in the "Chinese Ethnographic Dictionary",and its F1 value is as high as 0.8136.2.Constructing the knowledge graph of Chinese ethnic medicine: Use the constructed BERT-Bi LSTM-CRF knowledge model to extract the knowledge of the “DICTIONARY OF CHINESE ETHNIC MEDICINE”,and then construct the knowledge grahp model of Chinese ethnic medicine,which specifically refines 7 entities and 9 relationships.Finally,according to the model,the extracted knowledge was imported into Neo4 j,and the knowledge graph of Chinese ethnic medicine was established.3.In order to further promote the intelligent development of knowledge in the field of Chinese ethnic medicine,based on the constructed knowledge graph of Chinese ethnic medicine,the application platform of the knowledge graph of Chinese ethnic medicine was constructed based on web components,and the visualizatione and intelligent question and answer of Chinese ethnic medicine knowledge was explored;In daily life,users can use this platform to inquire about ethnic medicine related knowledge.