Research On Realization And Application Of Spatial Co-location Rule Mining Algorithm Based On GIS

Spatial Data Mining (SDM) is a very important branch of Data Mining. It has great efforts on understanding spatial data; find the intrinsic correlations among spatial data, and between spatial data and non-spatial data, and expressing the rules of spatial data concisely, which allow the extraction of implicit knowledge, spatial relations, or other patterns not explicitly stored in spatial databases.SDM face to spatial databases which are important and special. Geographic Information System (GIS) is the carrier of spatial databases with a mass of spatial data and attribute data. Therefore, using GIS as the framework of SDM tool can make the use of spatial data more conveniently and express knowledge more intuitively.Tobler's first law of geography describes the spatial correlations as: Everything is related to everything else but nearby things are more related than distant things. The discovery of spatial association rules is a descriptive mining task aiming at the detection of associations between reference objects and some task-relevant objects, the former being the main subject of the description while the latter being spatial objects that are relevant for the task at hand and spatially related to the former. If we can find the rules or mutual associations in these data, we can conclude the external world in reverse. This is the task of Spatial Association Rule Mining.General studies on spatial association rules are based on conventional association rule algorithms, which treat spatial databases as usual data sets. Co-location rule algorithms meet the demand of mining spatial association rules effectively and exactly.This paper consists of five chapters. Chapter 1 outlines the basic concepts, theories and applications of SDM, GIS and spatial association rules. Chapter 2 introduces basic concepts on spatial data and integration mode of GIS and SDM. Chapter 3 studies on spatial co-location rule algorithms. First we introduce the Event Center Model which deals with boolean spatial data, and the spatial co-location rule algorithm which deals with spatial multi- dimensional classified data. Then we propose an improvement on the latter by controlling the size of candidate co-location patterns effectively and reducing the scan time of database, in order to enhance efficiency of the algorithm. Chapter 4, take economy census statistical data in Daxing district, Beijing, 2005 as an example, analyze the spatial distribution rules of industry, catering industry and school, and obtain some conclusions using the improved algorithm. Finally, in Chapter 5, summarize the content, innovation and limitations of this paper.