Construction And Application Of Knowledge Graph In The Field Of Competitive Sports

With the rapid development of sports informatization and the wide application of artificial intelligence,intelligent informatization has played an important role in the field of competitive sports.At present,the amount of information data in the field of competitive sports is huge and scattered,the update speed is fast,and the sources are various.Therefore,in order to make management better and more convenient,the knowledge graph technology is put in the field of competitive sports.Establish the relationship between competitive sports entities,integrate these scattered data information,and store them in the graph database,and display the relationship between competitive sports knowledge in the form of graphs,so that knowledge query can be expanded and reasoned.According to the advantages of knowledge graph in the field of competitive sports,this paper mainly studies the construction method of knowledge graph in the field of competitive sports.Firstly,professionally define and classify the knowledge of competitive sports,organize the knowledge data in the field of competitive sports into a rule form that can meet the construction of knowledge graph,and then store the sorted knowledge into it through the Neo4 j graph database to construct a competitive sports knowledge graph.Finally,the constructed knowledge graph is used as the source of data to build a visualization system to make useful explorations for the application of knowledge graphs in the field of competitive sports.The specific research contents are as follows:(1)Ontology construction and definition annotation specification of competitive sports knowledge graph.Firstly,according to the professional knowledge and general knowledge of the competitive sports discipline,it divides and defines the types of the ontology concept and the relationship between the ontology in the competitive sports field,and then determines the model structure of the knowledge base in the competitive sports field.Then,according to the ontology classification and relationship classification The rules define and standardize the competitive sports data to prepare for the knowledge extraction experiment.(2)Extraction of competitive sports knowledge.The current mainstream and mature Bi LSTM+CRF named entity recognition model is used to complete entity extraction from multi-source competitive sports data,and then an improved BERT-based grid structure model BERT+Att+Bi GRU is introduced to extract entity relationships in the field of competitive sports.From the experimental results,it can be concluded that the BERT+Att+Bi GRU model can effectively improve the accuracy of entity relationship extraction by using explicit word information,and can better complete entity relationship extraction in the field of competitive sports.(3)Design and implementation of knowledge graph system in the field of competitive sports.Firstly,the prepared knowledge in the field of competitive sports is stored through the graph database Neo4 j,which improves the shortcomings of common relational databases for knowledge storage.Then make a demand analysis for the knowledge graph visualization platform system in the field of competitive sports,complete the structural design,use the Django-based framework to realize the functions of the platform system’s competitive sports knowledge query and competitive sports knowledge visualization,and verify the platform system through example tests.work performance.