The concept of safe region has been used to reduce the computation and communication cost for the continuous monitoring of k nearest neighbor (kNN) queries. A safe region is an area such that as long as a query remains in it, the set of its kNNs does not change. In this paper, we present an efficient technique to construct the safe region by using cheap RangeNN queries. We also extend our approach for dynamic datasets (the objects may appear or disappear from the dataset). Our proposed algorithm outperforms existing algorithms and scales better with the increase in k. © 2009 Springer Berlin Heidelberg.
CITATION STYLE
Hasan, M., Cheema, M. A., Lin, X., & Zhang, Y. (2009). Efficient construction of safe regions for moving kNN queries over dynamic datasets. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 5644 LNCS, pp. 373–379). https://doi.org/10.1007/978-3-642-02982-0_24
Mendeley helps you to discover research relevant for your work.