This study aims to analyze land use patterns using time-series satellite images of Seoul Metropolitan Area for the past 30 years, and present a macroscopic model for predicting future land use patterns using Markov Chain based probability model, and finally examine its applicability to Korea. Several Landsat MSS and TM images were used to acquire land-use change patterns and dynamic land-use change patterns were categorized from the classified images. Finally, spatio-temporal transition matrices were constructed from the classified images and applied them into a Markov Chain based model to predict land-use changes for the study area.

AOI(Attribute-Oriented Induction) is one of the data mining techniques with which phenomena of complicated spatial relationships can be expressed as specific rules by summarizing spatial or non-spatial data in accordance with super ordinate concepts. In addition to such AOI technique, if GIS(Geographic Information System) that has an advantage to solve space-related problems is combined, they can be used to expect the issues of urban growth. Combining the AOI technique and GIS, the study draws out spatial association rules focusing on a physical urban growth, and those rules are applied to urban growth models during the period from the 1960’s to the 1990’s. The results and analysis of the urban growth modes combined with data mining are compared with those of Clarke Keith’s UGM(Urban Growth Model).

Data mining refers to the extraction of knowledge from large amounts of data using pattern recognition, statistical methods, and artificial intelligence. The decision tree method, a well-known data mining technique, is used to extract decision rules because using it, we can understand the grounds of classification or prediction. The decision tree method thus can be used to analyze spatial data. There is an enormous amount of spatial data in GIS (Geographic Information System), which has been studied in modeling with these data. Spatial modeling as applied to a decision tree method can be carried out more effectively. In this paper, we extracted spatial rules using the decision tree method, and then applied them to urban growth modeling based on Cellular Automata (CA). An evaluation comparing the model using the decision tree method (the proposed model) with the standard UGM showed that the proposed model is more accurate.