Research On Keyword Search Based On Knowledge Graph In Federated RDF System

In recent years,with the rapid development of artificial intelligence and the wide application of knowledge graphs,traditional keyword query methods are gradually unable to meet the user's query requirements.People have put forward new requirements for knowledge retrieval technology and are eager to find a more intelligent way.The knowledge graph has a large semantic background knowledge,and the keyword query function that based on the knowledge graph can more accurately understand the user's query intent and return the corresponding search result instead of the keyword matching web page.In this paper,we mainly study how to implement keyword query on the knowledge graph under the federated RDF database.The keyword search research on the existing knowledge graph is mostly based on constructing a global index,and is not suitable for the keyword search problem of the knowledge graph stored in the federated RDF database.Based on the knowledge graph stored under the federated RDF database,this paper studies the keyword search problem.It mainly includes the following parts: First,the knowledge graph is divided into different data sources in the federated RDF database,and the schema graph auxiliary information is constructed in the offline phase according to the knowledge graph data abstraction,and each keyword is found in the schema graph by means of the full-text search.The corresponding candidate class vertex is then mapped to the schema graph.Second,for the schema graph after keyword mapping,the sub-graph division algorithm divides several sub-structure graphs containing all the different keywords,and converts each sub-structure graph into the formal query language SPARQL.Each set of SPARQL query statements returns the final query result through the SPARQL interface of the data source.The innovations of this paper are as follows:1.For the problem that the knowledge graph stored in the federated RDF database cannot download the data to construct the full-text index,a new query transformation method based on schema graph and keyword mapping is proposed.A query processing method based on subgraph partitioning and SPARQL construction is proposed.Finally,the schema graph auxiliary information is transformed into SPARQL query statement to obtain the query result.The theoretical analysis and experimental results show that the proposed method can return the results within the acceptable query response time,and the experimental results have higher accuracy,which satisfies the keyword search requirements of the knowledge graph in the federated RDF database.And the work has the promotion value and application prospects.