Studies On Query Processing Techniques Over Uncertain RDF Databases

Semantic Web as a data network is constantly collecting and organizing Web information. Therefore, related applications are faced with the challenges of efficient access to large-scale RDF (Resource Description Framework) data contained in the Semantic Web. On the other hand, the prevalence of errors and noises in modern scientific research methods and measurement techniques, as well as the data clutter owing to data integration makes RDF data uncertain. In the field of Semantic Web, uncertain RDF data query has aroused the considerable attention of academia and become a newly developed research focus. The purpose of this paper is to design efficient and accurate algorithm to resolve uncertain RDF data queries with high efficiency.To summarize, we make the following contributions.Build uncertain RDF data benchmark. Taking into account the fact that uncertain RDF data is not available when acadamics study uncertain RDF data.The dissertation constructed an uncertain RDF data benchmark, based on the extending of an existing deterministic benchmark.This benchmark could generate uncertain RDF data with arbitrary size, based on skewed distribution or uniform distribution.Propose a highly efficient algorithm for uncertain RDF graph query. This paper firstly established an index tree in favor of effectively conducting subgraph matching. Secondly, the paper carried out structural pruning and probabilistic pruning taking advantage of summary information of RDF graphs in the subgraph matching query process. Then, the thesis proposed a sampling method for filtering the result set.Achieve transitive inference capability for uncertain RDF data. Paper in advance encoded information for vertex set of uncertain RDF graph to avoid traversing entire graph in realizing subgraph matching. Thereby we achieve transitive inference capability by using the approximation algorithm obtaining the path set. This method reduces the time and space costs and improves the overall efficiency of the query.Design experiments to verify conclusions.Based on uncertain RDF data bench-mark raised in my paper, the query algorithm was measured performance. Through the performance comparison of experimental results, the paper selected optimal para-meters for the algorithm and evaluated the proposed algorithm performance based on benchmarks as a reference point. The results show that the algorithm is highly efficient and accurate.