Knowledge Graph Construction in Logistics Based on Multi-source Data Fusion

In the context of the information technology of logistics enterprises, the field of artificial intelligence empowers the logistics industry and brings revolutionary changes to it. The emergence of drones, self-driving cars and intelligent picking robots has dramatically changed the existing mode of warehousing, transport and logistics operations. Information technology based on big data mining and natural language processing has also made "data-driven logistics" a reality. In order to make full use of the rapidly growing big data in the logistics field and to solve the problem of efficiently querying the information of multiple logistics enterprises for comparison and extracting, integrating and querying multi-source data of logistics enterprises in the era of big data, the article combines the special vocabulary and special grammatical features in the logistics field. It proposes a method for extracting and constructing the entity relationship based on multi-source data fusion. The coremethod is to design and build stored procedures using the MySQL database to transform structured batches into triples. A dependency syntactic analysis model is designed, and a custom dictionary is added to extract triads from unstructured data to improve the efficiency and accuracy of traditional dependency syntactic analysis. Based on the triad extraction results, a knowledge graph in the logistics domain is constructed in the Neo4j graph database. The graph can be used to compare multiple logistics enterprises' queries, uncover suspicious points of logistics enterprises, expose enterprise risks, effectively improve the efficiency of logistics information queries, and provide an innovative idea for logistics enterprises' information construction.