1 / 17

SlimSS -tree: A New Tree Combined SS -tree With Slim-down Algorithm

SlimSS -tree: A New Tree Combined SS -tree With Slim-down Algorithm. Lifang Yang, Xianglin Huang, Rui Lv and Hui Lv are with Computer School, Communication University of China,Beijing , China. Mario Rodriguez Revollo School of Computer Science, UCSP. INDEX. INTRODUCTION

calum
Download Presentation

SlimSS -tree: A New Tree Combined SS -tree With Slim-down Algorithm

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. SlimSS-tree: A New Tree Combined SS-treeWith Slim-down Algorithm Lifang Yang, Xianglin Huang, RuiLv and HuiLvare with Computer School, Communication University of China,Beijing, China Mario Rodriguez Revollo School of Computer Science, UCSP

  2. INDEX • INTRODUCTION • SLIMSS-TREE INDEX STRUCTURE • The structure of SlimSS-tree • Insertion Algorithm • Advanced Slim-down Algorithm • Nearest Neighbor Queries • CONCLUSION Mario Rodriguez Revollo - San Pablo, CatholicUniversity

  3. INTRODUCTION Mario Rodriguez Revollo - San Pablo, CatholicUniversity

  4. In this paper, we focus on the R-tree family. The most successful variant of R-tree seems to be R*-tree which imports the idea of ‘Forced-Reinsert’ • Slim-down algorithm is used in Slim-tree. It could diminish the number of objects that fall within the intersection of two regions in the same level, sometimes even decrease the number of nodes in the tree. Mario Rodriguez Revollo - San Pablo, Catholic University

  5. SLIMSS-TREE INDEX STRUCTURE Mario Rodriguez Revollo - San Pablo, CatholicUniversity

  6. The structure of SlimSS-tree • SlimSS-tree is proposed based on the SS-tree, which uses the super spheres to manage the nodes instead of the super rectangles. • In our design, The SlimSS-tree mainly consists of intermediate nodes, leaf nodes and date nodes. Mario Rodriguez Revollo - San Pablo, CatholicUniversity

  7. The example of SlimSS-tree structure Mario Rodriguez Revollo - San Pablo, CatholicUniversity

  8. Insertion Algorithm • The insertion algorithm of SlimSS-tree is similar to the insertion algorithm of R-tree. • There are two main differences. Mario Rodriguez Revollo - San Pablo, CatholicUniversity

  9. Mario Rodriguez Revollo - San Pablo, Catholic University

  10. The insertion process of inserting one node N to the tree is similar to the feature vector insertion process. In the node insertion process, the entry whose centroid is nearest to the centroid of the inserted node N will be chosen for inserting the node. Mario Rodriguez Revollo - San Pablo, CatholicUniversity

  11. Advanced Slim-down Algorithm • After all feature vectors are inserted to the SlimSS-tree and the centroid and radius parameters of all entries has been adjusted, the advanced Slim-down algorithm which combins Slim-down with Reinsertion will be used to postprocess the tree to make it more tighter. Mario Rodriguez Revollo - San Pablo, Catholic University

  12. How the Slim-down algorithm works Mario Rodriguez Revollo - San Pablo, Catholic University

  13. Mario Rodriguez Revollo - San Pablo, CatholicUniversity

  14. Nearest Neighbor Queries • the KNN query NN(Q, k) selects the k indexed feature vectors which have the shortest distance from Q. • MINDIST algorithm is one of the most popular nearest neighbor search algorithms. The MINDIST (Minimum Distance) of a point Q(q1, q2,…, qd) in Euclidean space E(d) from a sphere S(O, r) in the same space Mario Rodriguez Revollo - San Pablo, CatholicUniversity

  15. CONCLUSION Mario Rodriguez Revollo - San Pablo, CatholicUniversity

  16. In this paper, we use the Slim-down algorithm combined with Reinsertion in SS-tree, and propose a new index structure: SlimSS-tree. • In the high dimensional space, the SlimSS-tree outperforms the SS-tree obviously, while in low dimensional space it almost has the same performance as the SS-tree. • Slim-down algorithm can effectively reduce the intersection regions between spheres, thus improve the search performance of SlimSS-tree obviously. Mario Rodriguez Revollo - San Pablo, CatholicUniversity

  17. SlimSS-tree: A New Tree Combined SS-treeWith Slim-down Algorithm Mario Rodriguez Revollo School of Computer Science, UCSP

More Related