Research On The Application Of Question-answering System Based On Knowledge Graph In The Field Of Government Dat

With the wide application and continuous development of Internet technology,information is growing explosively.The traditional government search engine gradually can not meet the needs of public servants to retrieve information.The returned information is too miscellaneous,and it takes a lot of time and manpower to read the returned information for the public servants to get the real answer.It can be seen that the traditional search engine has been unable to meet the needs of users.The rapid development of knowledge map promotes the development of Q&A system in the field of government data.Based on the above background,This paper studies and develops an e-government Knowledge Q&A system based on knowledge map.It uses web crawlers to obtain egovernment data from e-government websites,and uses web page parsing tools Xpath and Beautifulsoup to extract and parse the downloaded original web pages.According to the crawled e-government data,it constructs e-government node information and formulates the relationship between different nodes,Form a relatively complete knowledge map in the field of government affairs.Through consulting the related papers of knowledge mapping and experiencing the question answering system of other platforms,the requirements of the system are determined from the perspective of technical feasibility,social feasibility and economic feasibility.After completing the requirement analysis of the system,the system is designed briefly,and the hierarchical architecture is adopted to reduce the dependence between layers,The function modules of the system are defined.On the basis of the outline design,according to the development process of software engineering,the function modules are designed and coded in detail.The core function modules of the system mainly include: data acquisition module,entity data storage module,problem preprocessing module,problem query engine module and background management module.Finally,the whole system is tested,and the system is improved according to the test results.The main contents of this paper are as follows.1.Use scrapy crawler framework to crawl the government website on a large scale,process the data structurally,analyze the attribute of question answering entity and the relationship between entities in the field of government knowledge,and construct the knowledge map for the field of government knowledge.2.Analyze the problem,including word segmentation,part of speech tagging,removing invalid characters and other preprocessing operations,and design an entity recognition model based on Bert-bigru-CRF to improve the accuracy and efficiency of entity recognition,A modified Bert-CNN model is designed to recognize the intention of questions.3.Realize the intelligent Q&A system of government knowledge based on knowledge map.In order to help government officials work,the system is developed with B / S architecture which is suitable for the office environment of civil servants,and the searched government documents are displayed in the form of visualization.According to the different functional modules of the system,the system is divided into server-side and user front-end.Using Flask framework to achieve the functional modules of the background server,the front-end development technology uses Bootstrap framework,combined with element UI components to complete the frontend interface construction.