A Study on Method of Extracting High-Risk Accident Causes in Nagano Using Bayesian Network Analysis

Road accidents have been caused by many factors and finding the root cause is very important. In previous report, logistic regression analysis method was used to identify high-risk accident factors. It was possible to analyze the significant factors, but the method alone cannot identify the causal chain and root cause of accidents. This paper presents an analysis of 86,331 traffic accident (TA) data in Nagano prefecture, where eight accident factors with odds ratio > 3 considered to cause high-risk accidents were applied to build Bayesian Network (BN)s. The aim of this study is to study the method of extracting high-risk accident causes in Nagano using Bayesian Network analysis. As a result, 37 BN causal chains were obtained with all sequence orders lead to high-risk accidents, which caused serious injuries or deaths in Nagano. The results also confirmed that the highest-risk accident in Nagano to be single-vehicle collision accidents happened due to law violation such as over speeding at intersection and the high impact causing severe injury or death. Using information obtained through BN analysis, more focus on high-risk accidents with typical characteristics in Nagano can be given and countermeasures can be constructed for cases that needed more attention with logical and reasonable approach. © 2022. Society of Automotive Engineers of Japan, Inc. All Rights Reserved.