Fuzzy rule extraction from GIS data with a neural fuzzy system for decision making
Zheng Ding, Wolfgang Kainz · 1999
This study focuses on a new methodology to model a complex process of fuzzy rule acquisition from GIS data. The proposed method of neural fuzzy network described in this paper differs from the other methods in construction of the network by a simple representation of specific knowledge. Our goal is to use semantic If-Then rules instead of numerical values in the analytical models when uncertainty is involved in data. Such a method is expected to solve realistic problems in natural resource analysis and management. This paper discusses an adaptive neural network based on a fuzzy inference system that is able to learn fuzzy sets and fuzzy rules from data by the algorithm of steepest gradient descent. The fuzzy rules that are created by this approach can be very well interpreted to support decision-making. The method is tested in a case study of land suitability assessment. The result shows that the neural fuzzy network successfully extracted fuzzy rules with multidimensional spatial features obtained from GIS and simulated the decision process.