Top-k Closest Pair Queries over Spatial Knowledge Graph

Fangwei Wu, Xike Xie, Jieming Shi

Abstract: Recently, RDF data has been enriched with spatial semantics enabling spatial keyword search. Research spatial keyword search over spatial RDF data focus on finding the spatial entities rooted at subtrees which cover given query keywords. In this work, we study how relevant spatial entity pairs can be efficiently retrieved, where the relevance is determined according to both spatial distances and textual similarities. The retrieved top-k closest pairs are ranked and then returned to users for the interests of business intelligence and recommendation. We propose a branch-and-bound framework associated with effective lower and upper bound pruning techniques and early stopping conditions for efficiently retrieving relevant top-k closet pairs. The results demonstrate the high efficiency of our proposal compared to baseline solutions.