Research On The Construction Of Knowledge Graph For Smart Tourism

The rapid economic development has significantly improved people's material living standards and promoted the development of the tourism industry.Especially in the background of big data,it is extremely important for people to search tourism information through the Internet to provide tourists with comprehensive and reliable travel information.Knowledge graph as an important technology in the era of big data can effectively integrate and manage tourism data resources and improve the quality of tourism services.However,most of the existing tourism information is mainly based on the recommended travel strategies of businesses.The tourism data of various businesses is relatively scattered,which is not beneficial for tourists to obtain tourism information.To address the above problems,this paper integrates various types of tourism information to construct a tourist-oriented smart tourism knowledge graph based on the needs of tourists,practical application scenarios.The tourist-oriented smart tourism knowledge graph improves tourism resource management,and provides users with accurate tourism service information.At the same time,this paper also improves the relationship extraction in the knowledge graph that can obtain the fine-grained relationship between the entities in the knowledge graph.The main contribution of this paper can be summarized as follows:(1)From the perspective of tourists,we analyze the source of tourism data,determine the scope of the tourism field,and build the ontology for smart tourism accroding to the actual scenario in tourism.Based on this ontology,the entity and relationship in tourism data are integrated to build a knowledge graph for smart tourism using a variety of data sources such as Ctrip and Meituan and a series of knowledge graph construction technologies.(2)To facilitate the tourists' access to tourism information,a tourism graph visualization system for tourists is developed by Django technology,Echarts technology,Tag Cloud technology,KNN algorithm,and THULAC tools based on constructing tourism knowledge graph.This system can implement features such as entity identification,entity query,relationship query,attraction overview,and entity encyclopedia.(3)An entity-relationship extraction model of multi-granularity feature fusion is proposed to respond to the single feature size problem in existing entity-relationship extraction models.Firstly,this model extracts textual features using PCNN units with different size convolutional nuclei,preserving structural information while extracting multi-grained features.Secondly,in the information fusion layer,multi-granularity features are fused in three ways to get a better fusion mechanism.Finally,experiments on the public corpus demonstrate that the model of this paper outperforms the mainstream model on the PR curve.