원문정보
초록
영어
Given a site set P and an object set R, Bichroamtic RkNN query of site q∈P finds objects that take q as k nearest neighbors, which can been used to evaluate the influence of q on objects. Existing methods execute RkNN query by pruning strategies based on spatial indexes. For any change of object datasets (like moving objects), these methods need to compute RkNN again. The paper proposes a new algorithm based on K-order Voronoi diagrams. For a fixed site q, its RkNN region does not change whatever object datasets are updated or not. So we only search objects in Voronoi region of q. In this paper, we first give some propositions that provide the searching bounds. Then, BRKVD algorithm is proposed for RkNN query based on R-Tree, which supports the frequent changes of datasets and k. The experimental results show that the proposed algorithm performs the existing algorithms on efficiency.
목차
1. Introduction
2. Problem Definition
3. RkNN Query Algorithm based on K-order Voronoi Diagram (BRKVD)
3.1. The Properties of K-order Voronoi
3.2. BRKVD Algorithm
3.3. The Index Structure of the Algorithm
4. Performance Evaluation
5. Conclusions
Acknowledgments
References