Font Size: a A A

Research And Construction Of Mongolain-chinese Bilingual Knowledge Graph For Tourism

Posted on:2020-06-19Degree:MasterType:Thesis
Country:ChinaCandidate:X L GengFull Text:PDF
GTID:2428330596992281Subject:Computer technology
Abstract/Summary:PDF Full Text Request
Today,with the rapid development of transportation and the Internet,tourism has become an important part of recent national income,and the “tourism + Internet” model has emerged.The information of tourism resources has become more and more complex and more and more complicated.Although people can easily search for some resources on the Internet,there are more and more useless information,which greatly interferes with the retrieval of useful information by the public.Therefore,how to make good use of the huge and complex information and effectively extract useful information from it becomes very urgent.In view of the above problems,this thesis constructs a Mongolian-Chinese bilingual knowledge graph for tourism.The construction process mainly includes the following four parts:First,constructing the semantic framework of the tourism domain,and extracting and processing the information according to the semantic framework.The data sources of this article mainly include interactive encyclopedia,Baidu Encyclopedia,Sogou Encyclopedia,and Ctrip.Second,Perform data preprocessing and build a corpus.Use the open source tool NeuralCoref to replace the referential words in the text with the entity name for better word segmentation and labeling.The word segmentation operation is completed by the NLPIR tool,and the labeling method is BMEO,and the corpus establishment is completed at this stage.Then,For the marked text,the CRFs toolkit is used to complete the named entity recognition,and the convolutional neural network is used to complete the relationship extraction.According to the extracted entity and attribute relationship,a triplet for constructing a knowledge graph is formed.The attribute values extracted therefrom are merged to select the optimal attribute value.The triads are processed in Mongolian by querying professional books and using the professional Mongolian processing platform.Finally,According to the construction process of the seven-step method,the Protégé tool is used to complete the construction of the ontology of the tourism field,and then on the basis of the domain ontology,the construction of the Mongolian-Chinese bilingual knowledge graph for tourism is realized.
Keywords/Search Tags:Tourism resources, Knowledge extraction, Knowledge graph, Mongolian-Chinese bilingual, Named entity recognition, Convolutional neural network
PDF Full Text Request
Related items