Open AccessQuestion answering system based on tourism knowledge graph
Author(s) -
Yuan Sui
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1883/1/012064
Subject(s) - computer science , tourism , question answering , information retrieval , graph , knowledge graph , matching (statistics) , knowledge base , graph database , artificial intelligence , world wide web , theoretical computer science , geography , statistics , mathematics , archaeology
Nowadays tourism information services only provide users with massive and fragmented information returned by independent network search which makes users often need to spend a lot of time and energy to find what they really want from the massive data. As a result, route designing is very complicated. In view of this situation, this study builds a tourism knowledge graph based on neo4j and constructs a question answering system (QA). Also, we carry out the model and system performance evaluation, trying to improve the user satisfaction with query experience. According to the structure of question answering system (QA), this research designed and implemented named entity recognition (NER) model based on Bert-BiLSTM-CRF and matching reasoning model based on templates. With the above methods, natural language questions were successfully transformed into cypher query statements recognizable in graph database, and the corresponding answers will be captured and returned from tourism knowledge graph. According to the experiment, the method of Bert-BiLSTM-CRF obtains the state of art and QA system performs quickly and efficiently. For the purpose that artificial intelligence helps the development of tourism industry, this study has a certain significance.