다차원 검색공간에서 공간관계를 고려한 공간색인방법

Spatial Indexing Method Considering Spatial Relationships in Multidimensional Search Space

Reverse nearest neighbor queries frequently occur in several practical situations such as marketing-based profiles, decision making, resource management, image processing, and GIS. A spatial indexing method that considers spatial relationships is required to efficiently process the reverse nearest neighbor queries. In this study, a new index structure, R*nn-tree, which optimizes the processing cost of the reverse nearest neighbor queries is designed and implemented. R*nn-tree constructs an index by employing computation algorithms that use the degree of spatial relationships between the object and subspace. R*nn-treeoutperforms the existing methods regarding various aspects in both static and dynamic scenarios.