Building an expert system based on a geographic information system : an example of landuse management, Inner Mongolia, PRC

Abstract

This study was undertaken to develop a landuse expert system based on a geographic information system (GIS). More specifically, this study has developed a method of Knowledge Extraction from GIS (KEGIS) by machine-learning technique. This method has been tested by creating a landuse consulting system for Wongnute county, Inner Mongolia. The urgent need of resources management of Inner Mongolia is widely recognized: The ever-increasing population leads to pressures on the available resources that exceed their carrying capacity; The environmental collapse and resource degradation results in erosion and desertification. In addition, scarcity of landuse experts in Inner Mongolia limits the potential improvement for correct landuse. Artificial intelligence, especially expert system technology, offers the opportunity of transferring expertise analysis to this area. By developing a knowledge base that contains the relationships of landuse and environmental conditions, a landuse expert system can provide landuse suggestions which are normally performed by landuse experts. The difficulty of knowledge acquisition has limited the application of expert systems in geography. An alternative approach has been developed in this dissertation to overcome this obstacle. With this approach, geographic knowledge can be extracted from the information stored in a GIS. This study is an innovative breakthrough in the field of integration of AI with GIS. For testing the method, 154 sample-areas have been selected from Chi-Fong, Inner Mongolia, for knowledge base extraction. With the landuse-knowledge base, an inference engine, and a user interface, a landuse expert system has been constructed for landuse consulting. In an accuracy test, 19 of 25 cases (76 percent) are successful...