Spatial Data Mining With Fuzzy Decision Trees
Price
Free (open access)
Volume
22
Pages
14
Published
1998
Size
1,127 kb
Paper DOI
10.2495/DATA980171
Copyright
WIT Press
Author(s)
C. Marsala & N. Martini Bigolin
Abstract
Spatial Data Mining with Fuzzy Decision TVees C. Marsala & N. Martini Bigolin* LIP6, Universite Pierre et Marie Curie, ^ p/ace Jwsszew, 7&g&g f arzs cecfez 03, F/MTVCE. EMail: { Christophe.Marsala,Nara.Martini-Bigolin} @Hp6.fr Abstract In this paper, an approach is presented to search for useful patterns and dis- cover hidden information in Spatial Object-Oriented Databases (SOODB). Although many approaches of knowledge discovery for relational spatial da- tabases exist, there is a growing interest in mining SOODB. Indeed, object- oriented databases are well-suited to represent complex spatial information. Moreover, a very large number of existing spatial databases are ready to be mined. We propose an algorithm to mine a SOODB. After a spatial object query and a mathematical and fuzzy preprocessing, we apply decision tree based techniques and fuzzy set theory to discover knowledge. An experi- ment on a region of France to discover classification rules related to houses and urban
Keywords
