Urban Road Transport Network Analysis: Machine Learning and Social Network Approaches

Abstract

Traffic congestion is one of the most significant problems in urban transportation. It has been increasing, especially in regions close to intersections. Several methods have been developed to reduce the traffic congestion. One of the analysis methods is social network analysis (SNA). This method, which has increased use in transportation, can quickly identify the most central intersections in transportation networks. Improvements to central intersections, identified in a road network structure, speed up the traffic flow across the entire network structure. In this study, the Istanbul highway transportation network has been examined and values for a series of network centrality measures have been calculated using the SNA. The accuracy and error scales of the centrality values were compared using a machine learning algorithm. The Bonacich power centrality has been the best performance. Based on the study results the most central intersections in Istanbul have been determined.

Keywords

• central intersections
• machine learning
• transportation planning
• urban transportation