Evaluation of pedestrian transportation facilities in Taiwan using linear regression and support vector regression

• Published in: Road materials and pavement design Vol. 18; no. sup3; pp. 170 - 179
• Main Authors: Chen, Shih-Huang, Wu, Chia-Chen, Li, Pei-Yi, Adhitana Paramitha, Putri
• Format: Journal Article
• Language: English
• Published: Abingdon Taylor & Francis 10.07.2017; Lavoisier
• Subjects: Air temperature; Indication; Pedestrians; Population (statistical); Ratings; regression; Sidewalk; Support vector machines; SVR; Transportation; Urban areas; Walking; Walkways; Taiwan
• ISSN: 1468-0629; 2164-7402
• DOI: 10.1080/14680629.2017.1329872

Walking is the most ancient and basic form of transportation. Any travel, long and short, will involve some form of walking. As societies continue to develop and population keeps growing, the level of services provided to accommodate this most basic form of transportation has become an indicator of the development stage of a society. In urban areas, sidewalk is the most utilised transportation facility for walking; therefore, conditions of sidewalks are an indication of the level of services provided to pedestrians. As Taiwan moves towards becoming a developed region, providing safe and pleasant facilities for pedestrian transportation has been an important goal for the government. In this study, the performance of sidewalks, located throughout Taiwan, was assessed. A total of 196 sidewalks, selected from 22 cities and counties, were evaluated. An expert panel, consisting of 6-8 people, was assembled to assess the sidewalk performance. The ratings developed from the survey were defined as the Sidewalk Serviceability Index (SSI) and were used to represent the performance of the sidewalks. Eight sidewalk characteristics were also measured in the field. The eight characteristics were original width, net width, vertical clearance, vertical drop-off, ramp slope, transverse cross slope, illuminance, and air temperature. Correlations between the SSI and the measured sidewalk characteristics were developed using statistical techniques.