Research And Implementation Of Tourism Question And Answer System Based On Knowledge Graph

At present,with the rapid development of the Internet technology and the artificial intelligence technology,the mode of "internet + traditional industry" is playing an increasingly important role in various fields.Taking the tourism industry as an example,vigorously developing the "Internet + tourism" mode and innovating the tourism management by using advanced technologies such as Big Data and Deep learning.These measures have gradually played an important role in the development of tourism in China.In the past,when people visited scenic spots during the peak period of tourism,they needed to obtain information through manual consultation.The emergence of question answering system which applies natural language processing technology reduces unnecessary time consumption when people visit scenic spots.Compared with traditional information retrieval methods such as search engines,question answering system can return more concise and accurate answers to users,and improve the efficiency of information retrieval to a certain extent.Question answering system based on knowledge graph is an intelligent system that allows users to quickly find correct answers on the knowledge graph which has a large amount of structured knowledge.Taking Guangxi as an example,this paper presents a question answering system which combines the traditional FAQ(Frequently Asked Questions)model with.A fine-grained question answering over knowledge graph based on BiLSTM-CRF(Bi-directional Long Short-Term Memory-Conditional Random Field).The main research work of this paper is as follows:(1)Beacuse The scenic spot data collected on a single website is incomplete,this paper uses Scrapy framework in Python to crawl Guangxi scenic spot data from Ctrip.com,qunar.com and tuniu.com respectively,and then integrates the multi-source data into Neo4 J to construct Guangxi tourism knowledge map.(2)To solve the problem that the process of filtering candidate subjects in Question Answering over knowledge graph was too complex,and a majority of models ignored the fine-grained correlation between questions and relationships.A fine-grained question answering over knowledge graph based on BiLSTM-CRF is presented.In the entity detection part,BiLSTM-CRF is used to improve the accuracy of entity recognition.Then,the N-Gram model is combined with the Levenshtein Distance to avoid the complicate process in filtering candidate subjects;In the relation detection part,attention mechanism and CNN(Convolutional Neural Networks)are used to capture the correlation between questions and relationships from the semantic level and the word level.And evaluate our model on the datasets of FB2 M and FB5 M under FreeBase.Compared the accuracy of predicted entity-relation pairs with the existing methods..The results show that our method have been achieved state-of-the-art performance than before.(3)Based on the Guangxi tourism knowledge graph,this paper implements A fine-grained question answering over knowledge graph based on BiLSTM-CRF model.The Question answering system constructed combines the traditional FAQ Question answering model with the Question answering model based on knowledge graph,which can better meet the high requirements of tourists for information retrieval.