Researches Of Spatial Data Mining Based On Rough Set

The large number of spatial data, owing to rapid development of modern science and technology, and data retrieval devices, enable spatial data mining and knowledge discover become more important. In the same way, for information auto-collection, dynamic supervise manage and assistant decision-making to forestry resource, information technology will change the traditional forestry manage mode. This will certainly produce enormous original data. How to use such important data to serve forestry assistant decision-making is a significant problem.This paper first apply the rough set theory to forestry information manage. Site factors data base and sub-compartment attribute data are pretreated to reduce the dimensionality, this improve the efficiency of data analysis observably by such attribute reduction without losing original classification information. Spatial data mining also introduced data pretreatment, and detail method are provided. By spatial data pretreatment, original spatial data will be precise to reflect essence of problem, that are useful to retrieval more worthy rules in the data mining algorithms.This paper use MapInfo managing the spatial data, so spatial relations between spatial objects are implemented with MapX, include topology relation and direction relation auto-differentiation. After processing of spatial data, it can store in attribute data base. That are convenient to classical data mining technology for further analysis. In this way, we can make use of good methods and ideas in classical data mining to discover spatial knowledge under well-pleasing result.