Design And Implementation Of Medical Q&A System Based On Knowledge Graph

The rapid development of the Internet has brought great convenience to people’s lives.However,the quality of information on the network is mixed and increasing day by day,making it difficult for traditional search engines to provide accurate answers.Intelligent questionanswering systems can understand user questions and provide accurate,concise,and clear answers quickly,thus saving users time in searching for answers.Knowledge graph is one of the important cornerstones of artificial intelligence,which provides a high-quality and structured semantic knowledge base and strong data support for intelligent question-answering services.Due to the massive amount of medical information on the Internet,it is difficult for users to screen out useful information in a short time,which has become a universal problem.Therefore,this paper designs and implements a knowledge graph-based medical question-answering system on the basis of constructing a medical knowledge graph to solve the problem of user consultation needs.The main contents of this article include the following aspects:(1)The construction of medical knowledge graphs.Data is crucial in building a knowledge graph.The raw data is obtained from medical information websites using crawler technology,and the medical nodes and node relationships are created by data cleaning,formatting and defining entities,relationships and attributes.Finally,all data are saved to the Neo4 j graph database,thus successfully building a medical knowledge graph.(2)Question and answer system algorithm design.Regarding named entity recognition algorithms,this paper proposes the Ro BERTa-WWM-IDCNN-CRF algorithm model.The Ro BERTa-WWM pre-training model is introduced to improve the overall understanding of natural language,and IDCNN is used instead of Bi LSTM to improve the performance of the model,and a better Ro BERTa-WWM-IDCNN-CRF algorithm model is implemented.Bi LSTMCRF algorithm model by about 13%.Regarding the intent recognition algorithm,this paper proposes a Ro BERTa-WWM-Text CNN algorithm model based on Ro BERTa-WWM to parse the semantics of question sentences,and a Ro BERTa-WWM pre-training model is used to improve the overall understanding of natural language,which is transformed into a vector and passed into Text CNN,and then a Text CNN model is used to analyze the intent of the question sentences.(3)Design and implement of medical Question and Answer system based on knowledge graph.The Flask framework is used to integrate the constructed knowledge graph and various modules of the Question and Answer system on the back-end.The front-end pages are developed using Vue to provide users with a simple and practical communication interface,thus alleviating the medical pressure of user medical treatment.