Design And Implementation Of COVID-19 Question Answering System Based On Knowledge Graph

The high-speed information output of the Internet puts forward higher requirements for information retrieval technology,which prompts people to change from the traditional search engine retrieval system to a more friendly intelligent question-answering system when obtaining information.The emergence of Knowledge Graph makes it more convenient to store and understand structured knowledge data,which promotes the application of Knowledge Graph Question Answers in various fields.In the medical field,At the time of the global outbreak of COVID-19 epidemic in 2020,the demand for acquiring medicalrelated knowledge is prominent.In order to meet the needs of users for COVID-19 medical knowledge acquisition,a COVID-19 medical question answering system based on knowledge graph was designed and implemented.In terms of content,it mainly focuses on the medical field of COVID-19.Through the following three aspects of research work,a question-answering system based on knowledge graph is designed and implemented.Firstly,through the process of data collection and pretreatment,knowledge extraction and fusion,a COVID-19 medical knowledge graph was finally constructed and stored by the map database.The data source was to crawl the data of major encyclopedia websites and medical websites by using crawler.Then,combined with COVID-19 medical knowledge graph,a feature fusion question-answering method based on pre-training model was designed.This method adopted multi-layer feature fusion.For example,superficial features such as lexical and syntactic features and deep semantic features were integrated in the path sorting stage.Finally,Django,a Web framework developed based on Python,is adopted to implement the server side of the system,and "New Medical Answer" App is implemented based on Android,which is used as the main entrance of users.Meanwhile,in order to be compatible with PC users,a Web application platform is implemented.Thus,the COVID-19 medical question answering system based on knowledge graph is realized.Through experiments on manually labeled Chinese question-answering datasets,the F1 value of the proposed intelligent question-answering method reaches 90.8%,which is significantly higher than that of the question-answering method without a certain feature.In addition,the designed and implemented COVID-19 medical question-and-answer system can effectively provide users with efficient COVID-19 epidemic prevention and control,public opinion guidance,intelligent consultation and encyclopedia question-and-answer services.