Performance evaluation of tree-based structures

Abstract

Tree-based spatial indexing techniques have been developed for advanced applications which require high dimensional data. However, studies on spatial access methods particularly focus on the data structures and algorithms, performances in the worst case may not be predicted. In this paper, we propose a methodology for evaluating the performance of tree-based indexing techniques in the worst case, and a new tree-based data structure for indexing. The methodology allows better selection of indexing techniques for applications without having to actually implementing and experimenting the techniques. The new tree-based structure allows more efficient storage space utilisation and better searching time in comparison to the R-tree and TV-tree.