| With the fast development of network technology represented by 5g and artificial intelligence technology represented by deep learning,a new tourism management mode has been gradually formed by using new technologies like deep learning and big data in the tourism industry.Especially in terms of information acquisition,in the past,when people obtained tourism information,the way was mostly traditional search engine,which could not obtain information quickly,conveniently and accurately.The question answering system with natural language processing technology as the core can more accurately comprehend the user’s intention in a limited field,and feed back more accurate and concise answers,so as to greatly improve the speed and accuracy of users’ access to information.The knowledge base of question answering system is a direct factor affecting the effect of question answering system.Knowledge map has gradually become the main choice of knowledge base of question answering system because of its advantages in structured and relational data storage.China’s economic level is growing year by year,so people’s consumption level also began to grow,the number of outbound tourism has also increased.Australia is a popular destination for Chinese outbound tourism,but it is not very convenient to obtain tourism information about Australia.So,Based on the technology of named entity recognition,text classification and knowledge map,a question and answer system for Australian tourism knowledge map is developed.The main job of this paper is as follows:（1）The construction of knowledge map in the Australian tourism field.There are several core steps in the construction of knowledge graph,namely data acquisition,entity and relationship definition,entity disambiguation and data storage.First,crawl data from multiple travel websites through python’s request framework and xpath syntax.Secondly,define entities and extract relationships base on the characteristics of the data,and then use text similarity technology to complete entity alignment and attribute fusion of data from different sources,and convert the data into triple data format.Finally,according to the neo4 j graph database,a knowledge graph in the Australian tourism domain is constructed.（2）Design and implementation of key tasks of Q ＆ a system based on knowledge map.The key tasks of the question answering system based on knowledge map are further refined.From the process of answer search,the task is divided into two tasks:question entity recognition and question intention recognition.For the task of question entity recognition,a model made up of BERT-Bi LSTM-CRF is designed to extract question entities,and experiments are performed to verify that this method is better than other traditional methods.For the task of question sentence intent recognition,it is converted into a text classification task,and the BERT-softmax model is designed in combination with the data characteristics.The BERT model is used to realize the vector representation of the travel question and the softmax layer for classification,and through experiments to verify that this method is better than other The text classification model works better.（3）Design and implementation of Q ＆ a system based on knowledge graph of Australian tourism domain.Based on the process of Q ＆ a system based on knowledge map,taking knowledge map as knowledge base,question entity recognition model and question intention recognition model as semantic analysis method,a question answering system based on knowledge map in tourism field in Australia is designed and realized.The system can interact with users in real time.According to the questions raised by users in the field of tourism in Australia,the entity and intention categories are identified by using the question entity recognition model and question intention recognition model,and then search according to the entity and intention in the knowledge map and return accurate answers.To sum up,this paper designs and realizes a Chinese Q ＆ a system based on knowledge graph,and has done an exploration in the field of question answering system based on knowledge graph,which has certain theoretical meaning.Furthermore,this paper developed a Q ＆ a system based on the knowledge map of the Australian tourism field,and the test proved that the system has a good effect on the question and answer of relevant facts,and can quickly and accurately answer the user’s questions about Australia tourism,which satisfies the The needs of users have certain practical value and significance. |
