| With the development of Semantic Web,the scale of RDF data increases rapidly and semantic relationships become more complex,which makes the query efficiency of large-scale RDF data face unprecedented challenges.Since SPARQL queries can also be viewed as a query graph,retrieving information from large-scale RDF data using SPARQL can be viewed as a subgraph matching problem on a large graph.In order to improve the efficiency of large-scale RDF data query,this thesis uses the structure and semantic characteristics of SPARQL query graph,proposes a SPARQL subgraph pattern matching algorithm based on VF2,designs a RDF data query method based on R-VF2,and implements a large-scale RDF graph data query system.The main contributions of this thesis are as follows:(1)SPARQL subgraph pattern matching algorithm based on VF2.Aiming at the problem of low efficiency of subgraph pattern matching,this thesis proposes a SPARQL subgraph pattern matching framework combining the idea of VF2 subgraph matching algorithm and the characteristics of different subgraph patterns in SPARQL query.The framework realizes the type recognition of SPARQL query subgraph pattern,designs chain and star matching rules based on the pattern characteristics of chain and star subgraphs,and implements SPARQL subgraph pattern matching algorithm R-VF2.The algorithm compares three existing methods on four sizes of RDF data sets.Experimental results show that the space storage consumption of data sets is reduced by more than 30%.For the chain query graph,the query speed of the proposed algorithm is improved by 34,6.25,5.5 and 12.5 times on average under different data sizes,while for the star query graph,the query speed is improved by 54.5,11.5,8 and 24.5 times on average.The thesis also compares the effects of star matching rules and chain matching rules on query efficiency.The data show that in the same star query graph,compared with the chain matching rule,the star matching rule improves the query efficiency by 21.75%,indicating that the design of star matching rule has a better advantage for the query of star structure.(2)RDF data query method based on R-VF2.To solve the problem of SPARQL complex query on RDF graph,this thesis proposes an RDF data query method based on R-VF2 algorithm.This thesis analyzes the basic properties of SPARQL complex query graph,mines the topological structure characteristics and semantic information of complex query graph,proposes the decomposition strategy of SPARQL complex query graph,designs and implements the decomposition algorithm of complex query,and uses R-VF2 subgraph matching algorithm to realize the matching of SPARQL query decomposition subgraph.On this basis,according to the original basic characteristics of SPARQL query graph,the connection calculation of subgraph is realized,and the SPARQL complex query result is finally obtained.For complex query graph,the thesis compares the query efficiency of RDF data query method with SHARD,Pig SPARQL and S2 X methods under different sizes of RDF data sets.Experimental results show that the query speed of RDF data proposed in this thesis is improved by 58.5,6,7.25 and 31.5 times on average.(3)Large-scale RDF graph data query system.In order to provide a SPARQL query platform for users,this thesis combines the SPARQL subgraph pattern matching algorithm based on VF2 and RDF data query method based on R-VF2 to build a large-scale RDF graph data query system.The system implements RDF data processing,SPARQL query processing and matching module on RDF graph.The system application results show that the system can be applied to the academic retrieval field,and can help users to achieve visual effects in SPARQL query,so as to obtain better user experience. |
