文摘
Searches for objects associated with location information and non-spatial attributes have increased significantly over the years. To address this need, a top-k query may be issued by taking into account both the location information and non-spatial attributes. This paper focuses on a distance-based top-k query which retrieves the best objects based on distance from candidate objects to a query point as well as other non-spatial attributes. In this paper, we propose a new index structure and query processing algorithms for distance-based top-k queries. This new index, called SKY R-tree, drives on the strengths of R-tree and Skyline algorithm to efficiently prune the search space by exploring both the spatial proximity and non-spatial attributes. Moreover, we propose a variant of SKY R-tree, called S2KY R-tree which incorporates a similarity measure of non-spatial attributes. We demonstrate, through extensive experimentation, that our proposals perform very well in terms of I/O costs and CPU time.