(A) refinement strategy for selecting spatial data with polygonal query window

Abstract

The shape of query windows for spatial selection queries is a rectangle in many cases. However, it can be issued for spatial selection queries with not only rectangular query widow, but also polygonal query window. Moreover, as the more spatial data the applications like GIS can manage, it can support the more various applications. Therefore it is valuable for considering about the refinement processing method suitable for not only rectangle, but also general polygonal query window. It is the general state-of-the-art approach to use the plane-sweep algorithm for the computation algorithm in the refinement step as the spatial join queries, because the shape of spatial data and query window are all the general polygons. However, from the observation on the characteristics of spatial data and query windows, we can find in many cases that the shape of query window is much simpler than that of spatial data. Therefore there is something to be improved if we use the state-of-the-art approach, which is suitable for the two complex polygons, as the refinement processing strategy in this some situation. From these observations, we suggest a new refinement process approach using PBST (Partition Based Sequential Testing) based on PEC (Partial Enclosed Circle) we developed. And we evaluated the performance of this new approach and the state-of-the-art approach by experiments with the real map data. From these experiments, we found that the new approach using PBST is more suitable than the other one as far as its query window is composed of up to 20 vertices. Although there is some deviation in the experimental result according to the characteristic of the real map data, if the number of vertices composing the query window is less than about 20, the new approach using PBST we suggest is superior to the state-of-the-art approach using Plane-Sweep by about 20% in the cases of small queries ranging about 1% which the most parts of spatial selection queries are issued in the r...