Extracting features of sidewalk space using the rough sets approach

The aim of this paper is to offer a contribution to the study of the relationship between the physical elements of a sidewalk and people’s perception of it by applying the rough sets approach. In previous studies in this context statistical methods such as variance analysis and regression analysis are the most used techniques, and these suffer from the limitations of large sample sizes and rigorous statistical assumptions. However, the rough sets approach, as a nonparametric method of artificial intelligence, can handle this problem. In this study, twenty civil-engineering students and twenty noncivil-engineering students took part in an image preference survey assessing photographs of sidewalks. Then the rough sets approach was conducted by a four-step procedure to analyze the survey results. Finally, the output of the rough sets approach is used to express people’s perception of good and bad features of sidewalk space. The results will hopefully help planners to make effective decisions about sidewalk features.