Conference Paper

Nearest Neighbor Queries for R-Trees: Why Not Bottom-Up?

DOI: 10.1007/11733836_68 Conference: Database Systems for Advanced Applications, 11th International Conference, DASFAA 2006, Singapore, April 12-15, 2006, Proceedings
Source: DBLP


Given a query point q, finding the nearest neighbor (NN) object is one of the most important problem in computer science. In this paper, a bottom-up
search algorithm for processing NN query in R-trees is presented. An additional data structure, hash, is introduced to increase
the pruning capability of the proposed algorithm. Based on hash, whole data space is disjointly partitioned into n × n cells. Each cell contains the pointers of leaf nodes which intersect with the cell. The experiment shows that the proposed
approach outperforms the existing NN search algorithms including the BFS algorithm which is known as I/O optimal algorithm.

5 Reads