Analysis of Major Road Traffic Accident Causes Using a Combined Method of Association Rule and Complex Network
Shuai Huang, Cheng Jin, Tao Chen, Zhengwu Wang, Jie Wang
Abstract
To identify the key causes of major road traffic accidents (resulting in three or more deaths), this study constructed an accident causation network based on association rules and complex networks using data from 173 major traffic accidents over the past decade. Initially, 62 potential risk factors were extracted from aspects such as human, vehicle, road environment, and management. Association rule algorithms were then employed to explore the coupling relationships between these risk factors, generating strong association rules. Finally, a complex network model was built based on these association rules to identify critical risk factors. Results indicate that (1) 80% of major traffic accidents are linked to poor driving behavior. Complex network analysis identified speeding, overloading, lane crossing, and failure to maintain safe following distances as primary human factors, which are closely related to vehicle, road, and environmental conditions, contributing collectively to accidents. (2) Association rule results revealed that head‐on collisions are primarily related to lane crossing and occur on national and secondary grade highway; falling accidents are common on roads with inadequate infrastructure; rear‐end collision often happen on expressway and national highway, especially at night (18:00–07:00) with freight vehicles; vehicle‐related major accidents are usually associated with noncompliant vehicle performance, vehicle failures, overloading, and insufficient regulation. Major accidents involving inadequate road signs, markings, and safety barriers have a significant association with nighttime periods. The conclusions of this study can provide valuable insights for the prevention of major road traffic accidents.