Abstract
Previous studies on urban walkability have not identified the elements of attractive walking space in
suburban residential areas, nor the relationships among these elements. In addition, the evaluation of
walkability is based on individual experience and sensitivity, and is therefore uncertain. The use of Bayesian
networks is considered effective in modeling such an uncertain subject. The purpose of this paper is to discover
the components of space that residents of suburban residential areas find attractive when walking, and to clarify
the relationships among these components by creating a model using a Bayesian network.
