Spatial Databases and Indexing

Group Seminar - "Spatial Database and Indexing" by Murat Seckin Ayhan (ACTR 116, from Apr 18, 2008 09:00 AM to Apr 18, 2008 10:30 AM)

Abstract: Systems concerning of spatial data types have been frequently deployed, especially in Geographical Information Systems and applications. Various spatial access methods have been proposed for efficient query processing upon these complex and huge spatial data sets. In this study, some of these spatial access methods, R-tree, STR-tree, MX-CIF Quadtree and Gist R-tree, have been discussed and their fundamental window query performances on different data models have been focused. The performance results have showed that STR-tree has best query responses both in retrieval times and candidate sets whereas MX-CIF Quadtree stays behind R-tree as the data sets grow in size. Additionally, admirable performance gains, up-to 7-8 times, has been presented by built-in spatial index structure of PostgreSQL DBMS which is Gist R-tree.

slides

You may need powerpoint 2007 to open the file