Top-k Closest Pair Queries Over Spatial Knowledge Graph

Recently,RDF data has been enriched with spatial semantics(such as latitude and longitude and timestamps).RDF data are traditionally accessed using structured query languages,such as SPARQL.However,this requires users to understand the language as well as the RDF schema.Therefore,a new spatial keyword search model has emerged.Research spatial keyword search over spatial RDF data focuses on finding the spatial entities rooted at subtrees that cover given query keywords to ensure the relevance between spatial entities and query keywords,thus finding the spatial entity required by the user.On the basis of this model,by two sets of keywords,we can search the nearest spatial entity pairs on RDF data.Due to the large number of entities and rich text content in RDF data,all entities can not be quickly traversed within an effective time by normal join algorithms.In this work,we study how to establish indexes for RDF data,which are used to speed up the keyword matching algorithm,the nearest neighbor query algorithm,and the top-k ranking algorithm,whose content is as follows:(1)The dissertation purposes top-k closest pair queries over spatial knowledge graph,and the looseness between the spatial entity pair and query keywords.By combining the distance between the spatial entity pair and looseness,the dissertation proposes a ranking function for sorting multiple results.(2)The dissertation proposes a branch-and-bound framework associated with efficient lower and upper bound pruning techniques and early stopping conditions for efficiently retrieving relevant top-k closet pairs.(3)Based on the real data sets,the dissertation shows the effect of different parameters on the algorithms through experimental analysis.The comparison experiment shows that the optimization strategies can improve the efficiency of the algorithm.The results demonstrate the high efficiency of our proposal compared to baseline solutions.