Spatial Database Oriented Application Research On Spatial Data Mining

With China entering WTO and the arrival of commercial era of global cooperation, over 90% of decision-making is related to geographic data. In this age of Information, decision-making depends greatly on knowledge, but the relatively rare means of gaining knowledge limits the integrative utilization of datasets. The wide use of Geographic Information System (GIS), and the highly development of Data Mining, data collection of spatial data and spatial databases, as well as the disadvantage of the spatial analysis function, has resulted in the birth of Spatial Data Mining (SDM), a technique that scans and finds desired knowledge from spatial databases.Spatial databases is an aggregate of spatial data and attributive data, which describes spatial objects, and the database management system of collecting, storing, managing, searching, analyzing, expressing spatial data and attributive data.SDM, also known as Spatial Knowledge Discovery from Database, is a method of distilling interested spatial patterns and characters, prevalent relations of spatial data and no-spatial data, as well as other connotative characters of datasets from spatial databases. It is an extension of data mining in the appliance of Spatial Databases, which promotes GIS to become more intelligent and integrated.This thesis consists of six chapters.Chapter I introduces the background of Data Mining, summarizing the birth, the current research and the direction of its development, which indicates the importance and direction of the research of this thesis.Chapter II explains the basic theory of GIS and Spatial Databases. We explain the basic function and characters of GIS, spatial data model, spatial relationship, spatial query language and spatial index technique of Spatial Databases in detail.Chapter III begins with the process of SDM and the kinds of knowledge that can be found in Spatial Databases and goes on with the comparison between SDM and traditional Data Mining, showing their sameness and the differences in the angle ofprocessing and the techniques. After that, the techniques of SDM based on spatial data structure and other methods of SDM are explained. Then, we introduce several well-known prototype systems of SDM. Lastly, we study the application of SDM in the field of Spatial Intelligent Decision Support System (SIDSS) deeply, present three kinds of architecture of SIDSS based on SDM and go into detail the process of problem resolving of SIDSS based on SDM and Case-based Reasoning (CBR).Chapter IV starts with a discussion of the basic theory and the model of its network of forward Artificial Neuron Network (ANN). Then we briefly explain the architecture of its network and the design process of Back Propagation Algorithm (BP). After that, we discuss the main theory of Alternative Covering Algorithm. According to the peculiarity of spatial grid data, we apply Alternative Covering Algorithm to the classification of spatial data and make experiments for it. At last, we deeply analyze the results of the experiments. It shows that the algorithm we provide is of novelty and it can process very large amount of spatial data which consists of multiple attributive variables.Chapter V deeply explains research in the application of SDM. Firstly, we introduce the concept of logistics. Secondly, we design and implement a logistics prototype system based on GIS, briefly explain the architecture and function of the system, deeply analyze the key techniques, obviously provide the measures of integrating GIS and logistics system. Lastly, we apply some techniques of SDM to the logistics system based on GIS, which can promote the logistics decision-making to become visible and intelligent.Chapter VI serves as a summary of this thesis, pointing out future research directions in a prospecting point of view.