Font Size: a A A

Research And Application Of Semantic Approximate Top-k Query Over RDF Knowledge Graph

Posted on:2020-11-11Degree:MasterType:Thesis
Country:ChinaCandidate:Z P GeFull Text:PDF
GTID:2428330605966664Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
Knowledge graph is an important solution for large-scale information query and processing,it is a hot research topic.However,since some source sites of knowledge graph open collaborative editing entrances to users all over the world,the knowledge graph presents the characteristics of large data size,heterogeneity and diversity,which mains there are different forms of expression to describe the same knowledge,that is a huge challenge to achieve reliable and efficient knowledge graph query.Most of existing research work models knowledge graph query into sub-graph matching problem.There are two ways to solve this problem: one is sub-graph exact matching query,the other one is sub-graph similar matching query.Due to the large amount of noise data of knowledge graph,the first way is easy to lose the result that meets user's query intention.As for the second way,this technique usually generalizes query according to the structure of the subgraph without considering the semantic information implied in the graph,while the same relationship in the knowledge graph may have multiple semantic descriptions,which may result in the loss of qualified results for such queries.Therefore,it is necessary to design a method to obtain the semantic information implied in the knowledge graph,and propose a new query algorithm using the semantic information to improve the shortcomings of existing algorithm in the accuracy and efficiency of knowledge graph query.RDF(Resource Description Framework)is a storage form of knowledge graph,in order to improve the query accuracy and performance bottleneck of large-scale RDF knowledge graph approximate query under data heterogeneity and diversity constraints,we propose a semantic approximate Top-k query over RDF knowledge graph.We do research and optimization from the aspects of knowledge graph storage,knowledge graph semantic acquisition and knowledge graph query model.The main research work is as follows:(1)RDF knowledge graph data processing and storage optimization.We extract and preprocess the basic dataset,remove the redundant invalid information in the knowledge graph,and design the structure based on the adjacency table index to store the knowledge graph to meet the subsequent semantic query requirements.(2)We propose a local subgraph partitioning method and a local related corpus text acquisition method,transform the knowledge graph into corpus text,and use the mainstream text embedding model to train it to obtain the semantic vector.(3)We propose a top-k algorithm based on dynamic bounds to implement efficient semantic approximate query of RDF knowledge graph,which can ensure query efficiency while satisfying query precision.Based on above research results,we designed and developed the RDF knowledge graph semantic approximate query system for open dataset DBpedia,and verified the validity and usability of the research results.The research results of this paper will help to improve the query precision and solve the performance bottleneck problem of large-scale RDF knowledge graph approximate query under data heterogeneity and diversity constraints,it has important theoretical and practical significance for promoting the development of knowledge graph approximate query field.
Keywords/Search Tags:Knowledge Graph, Top-k Query, Semantic Approximate Query, Dynamic Boundary
PDF Full Text Request
Related items