摘要
提出路网中的互近邻查询问题.给定路网G(V,E),对象集P,查询点q,近邻数k1和k2,互近邻查询返回既是q的k1近邻,又是q的反k2近邻的对象集.为解决该问题,首先提出基础算法,即先求出查询点q的k1近邻作为候选,再验证这些候选是否为真正的结果.然后,在此基础上提出了优化算法,根据落在对象点与查询点最短路径边上的标记点个数直接排除掉一些错误的候选对象.最后,通过实验验证了优化算法的有效性.
The problem of mutual nearest neighbor(MNN) query in road networks is proposed in this paper. Given road networks G ( V,E), a set P of objects, a query object q, two parameters k1 and k2 , an MNN query returns from P, the set of objects that are among the k1 nearest neighbors of q; meanwhile, have q as one of their k1 nearest neighbors. To address this problem, firstly, the paper provides a fundamental algorithm that retrieves k1 nearest neighbors of q as candidate objects and then removes the false misses. Then, based on it, an optimal algorithm is proposed, which can remove some false misses directly according to the count of mark points which fall on the edge of shortest path between object and query. Finally, the experimental results show that the optimal algorithm is efficient and effective.
出处
《小型微型计算机系统》
CSCD
北大核心
2011年第4期606-610,共5页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(60773100)资助
关键词
互近邻
K近邻
反k近邻
路网
MNN
k nearest neighbor
reverse k nearest neighbor
road networks