Design And Implementation Of Question Answering System Based On Tourism Knowledge Graph

With the development of social economy,our country has been basically realize the comprehensive well-off,the people after meet the basic material needs,the demand for cultural products and services presents the fast growth the tendency,people will put more energy into leisure,leisure industry has become China's economy to a new economic growth point,There are more and more people take tourism as a way of leisure,but the satisfaction of tourism is often unsatisfactory.From the perspective of users,the cost of tourism consulting is high,and it is difficult to achieve professional,free and timely consulting services.From the perspective of tourism associations,the scale and level of tourism associations are not homogeneous,with limited knowledge of scenic spots throughout the country,and the tourism problems are so strange that they cannot even be answered.Therefore,in order to provide travelers with fast,free and accurate travel consultation and QA services to improve their travel experience and reduce the burden of travel associations,this system has a strong application value and practical significance.This thesis makes an in-depth study on the tourism knowledge graph.Aiming at the disadvantages of the difficulty of obtaining the desired tourist location,difficult to answer complex questions and reasoning,this paper designs and realizes the question and answer system of the tourism knowledge map.The system is mainly divided into the building module of tourism knowledge map,the problem query module and the knowledge map data update module.This system is based on the open source knowl-edge map clinga,supplemented by the open source knowledge map CASIA and the tourism data of the third-party tourism website ”China Tourism Network”.After ontol-ogy construction,knowledge acquisition,knowledge extraction and knowledge fusion,a tourism knowledge map is constructed.In order to identify the entities in the knowl-edge graph,the system mainly uses the dictionary to identify the existing entities.Due to the certainty of the dictionary,the recognition accuracy is very high.To the problem of natural language input of tourism after dictionary to identify the entity,through the entity attribute definition and template rule set again,use the template + triples in the form of dependency tree parsing,at the same time the system defines the article 220 of the most basic template queries,these template queries can complete queries on a single triple,the complex question is transformed into several simple questions by a triple dependency tree,and then queries are conducted on the constructed knowledge graph by calling the template of simple questions one by one,so as to complete the intelligentization of tourism recommendation question and answer.This paper provides a professional travel question and answer system for users.The user accesses this system through the browser,carries on the travel question to inquire.Before this,the system has established the knowledge graph of more than200,000 pieces of knowledge,and supports users to submit feedback and scenic spots to report,so as to carry out data expansion.Through the media of knowledge graph,the system combines the knowledge in the tourism field with the natural language problems input by users.Finally,the system proves that the system plays a helpful role in the travel recommendation of people with tourism intention through the analysis of real use cases.