Approach for Spatial Database Mining

Abstract

Most of the previous spatial mining works are depend on strategy of organizing the huge spatial data in a suitable data structure and usually the data organized as R-Tree. The data mining algorithms then applied on each level of R-Tree. This method causes time consuming and takes huge storage area and leads to inadequate results. The proposed approach suggests the following strategy for efficient spatial mining. It collects all the spatial data and organizes it (according to normalization and generalization) to a flat data base. After that the following steps will be executed: build the proposed spatial database, apply mining algorithms on the proposed Structure of the spatial data to extract the association rules, clusters and classes. Finally analyzes the resulted patterns from the mining algorithms.