Mountain roads,vital for linking remote areas and boosting regional development in China,suffer from poor conditions,complex environments,frequent accidents and severe consequences.Hence,establishing a predictive mode...Mountain roads,vital for linking remote areas and boosting regional development in China,suffer from poor conditions,complex environments,frequent accidents and severe consequences.Hence,establishing a predictive model for high-risk accident spots tailored to mountainous traffic conditions is crucial for formulating targeted preventive measures.Based on mountainous road accident cases in Guilin City,China from 2016 to 2020,this study employs spatial analysis methods to classify the data into high-risk and low-risk categories,determines key factors through Chi-square analysis and random forest Gini index and constructs an accident high-risk point prediction model using Bayesian networks.The results indicate that 19 variables,including gender,transportation mode and road location,are significantly correlated with accident risk status.Road type and traffic signal mode emerged as key factors directly influencing the overall risk profile,while other factors collectively affect accident risk status through indirect interactions.The highest risk scenario involves a male driver operating a passenger vehicle during the afternoon hours on a straight motorway in a hilly township area.The road is paved with asphalt and in good condition,equipped with basic safety measures such as signs and markings in the absence of signals.However,it suffers from insufficient traffic separation facilities,and there is a history of side-impact collisions causing injuries in this location.The predictive model developed in this study enables the forecasting of risk probabilities and the assessment of injury levels across different accident scenarios,providing a decision-making basis for formulating accident prevention policies for mountainous roads.展开更多
基金supported by the Natural Science Foundation of Guangxi Zhuang Autonomous Region(Grant No.:2023GXNSFAA026359)the Scientific Research and Technology Development Plan Project of Guilin(Grant No.:20230120-7)+1 种基金Additional support was provided by the National Natural Science Foundation of China(Grant No.:52262047)the Key Research and Development Program of Guangxi(Grant Nos.:AB25069483 and AB25069333).
文摘Mountain roads,vital for linking remote areas and boosting regional development in China,suffer from poor conditions,complex environments,frequent accidents and severe consequences.Hence,establishing a predictive model for high-risk accident spots tailored to mountainous traffic conditions is crucial for formulating targeted preventive measures.Based on mountainous road accident cases in Guilin City,China from 2016 to 2020,this study employs spatial analysis methods to classify the data into high-risk and low-risk categories,determines key factors through Chi-square analysis and random forest Gini index and constructs an accident high-risk point prediction model using Bayesian networks.The results indicate that 19 variables,including gender,transportation mode and road location,are significantly correlated with accident risk status.Road type and traffic signal mode emerged as key factors directly influencing the overall risk profile,while other factors collectively affect accident risk status through indirect interactions.The highest risk scenario involves a male driver operating a passenger vehicle during the afternoon hours on a straight motorway in a hilly township area.The road is paved with asphalt and in good condition,equipped with basic safety measures such as signs and markings in the absence of signals.However,it suffers from insufficient traffic separation facilities,and there is a history of side-impact collisions causing injuries in this location.The predictive model developed in this study enables the forecasting of risk probabilities and the assessment of injury levels across different accident scenarios,providing a decision-making basis for formulating accident prevention policies for mountainous roads.
