Research On Applications Of Clustering Analysis In GIS

There are abundant data and information in database of Geographic Information System (GIS), which include much implicit and valuable knowledge. Currently GIS is limited in collection, query and statistic of data. It can not discover information among data. By introducing data mining technologies into GIS, it helps to find those relationships and rules behind the data and serve the decision support. Recently, it has been an important research aspect known as Spatial Data Mining (SDM) which is used to mine spatial data objects.First, this dissertation proposes a rapid clustering algorithm for the spatial data in GIS based on density. It can handle spatial attributes and non-spatial attributes at the same time, find any-shape cluster, and identify the isolated points. Due to different scanning strategy, there is no necessary for the algorithm to have spatial query towards every points, which results in great time cost saving, and increases the clustering speed.Secondly, this dissertation also gives a mining algorithm based on Apriori. It's used to mine the clustering results and export association rules between them for the reference of making decisions. With only one scan of dataset, and without producing invalid candidate itemsets, it improves the efficiency of the Apriori algorithm.Finally, this dissertation puts forward an application framework with integration of those two algorithms. It meets the shortfall of the functions in GIS, and meanwhile plays powerful graphics functions of GIS. It also has been tested in a coal mine safety supervision and provides a possible solution for the integration of data mining and GIS.