Abstract
Roadside land use is one of the major factors which influence a long-term urban form and environmental quality. This study aims to develop a practical method for analyzing roadside land use pattern and its change based on a neural network model. Neural network model is regarded as a powerful method to describe complicated phenomena with a high non-linearity, however, it is often utilized as a black box, without logical interpretation. This study tries to overcome the crucial shortcoming existing in neural network by adding a detecting function of controversial and redundant description and to enlarge the applicability of neural network to land use analysis.
