Utilizing Partial Least Squares Path Modeling to Analyze Crash Risk Contributing Factors for Shanghai Urban Expressway System

ABSTRACT:The urban expressway systems are playing key roles in the metropolitan transportation system. However, frequent crash occurrences have significantly influenced the traffic operations and travel reliability. It is vital to understand the crash occurrence mechanisms and further improve traffic safety. Emerging studies have been conducted to unveil the relationships between crash risk contributing factors and crash outcomes with the advanced traffic sensing data. However, current results could mainly shed some lights on the correlation effects between traffic flow parameters and crash occurrence. In this study, the authors aimed at analyzing the confounding impacts of crash risk contributing factors and their causal relationships with crash occurrence through Partial Least Squares (PLS) Path Modeling approach. Crash data and traffic data from Shanghai urban expressway system were utilized. Firstly, potential crash risk contributing factors were summarized based on the literatures. Then, Random Forest (RF) model was adopted to rank the variable importance, and a total of six contributing factors were selected and used as inputs entered the PLS Path Modeling development procedure. Finally, the best PLS Path Modeling structures were identified, and crash occurrence scenarios and turbulent impacts on traffic flow parameters were concluded based on the analysis results.

Rongjie Yu, Yin Zheng*, Xuesong Wang, Zhen Gao. Utilizing Partial Least Squares Path Modeling to Analyze Crash Risk Contributing Factors for Shanghai Urban Expressway System. Transportation Research Board 97th Annual Meeting, Washington D.C., USA, 2018. 1.7-11.