山区农村公路交通事故差异化防控策略研究
Research on Differentiated Prevention and Control Strategies for Traffic Accidents on Rural Roads in Mountainous Areas

为探究山区农村公路交通事故致因，制定差异化安全防控策略，本文基于2016年至2020年广西桂林市农村公路事故数据，利用核密度方法识别得到6个事故高发区域；在此基础上，采用随机森林模型算法对事故高发区域进行影响因素分析，并运用十折交叉验证方法对分析结果进行模型检验。研究结果表明：(1) 在7点、10点、17~18点、21点时间段附近，事故的发生率相对较高；(2) 以一定角度碰撞和侧面碰撞的事故发生频率显著高于其他类型；(3) 相较于其他类型的道路，农村自建路发生事故的概率明显偏高；(4) 在能见度100~200米的道路状况下，更易诱发交通事故；(5) 没有标线的机非混合道是事故的高发区域。研究结果可为制定广西农村公路交通事故防控策略提供参考和依据，为我国农村公路安全提供了一定的参考价值。
To investigate the underlying causes of rural road traffic accidents in mountainous regions and develop targeted safety prevention and control measures, this paper analyzes rural road accident data from Guilin City, Guangxi, covering the period from 2016 to 2020. By applying the kernel density estimation method, six high-accident areas are identified. Building upon this foundation, the study employs a random forest model to examine the factors influencing these high-risk zones. The model’s performance is evaluated using ten-fold cross-validation to ensure robust analysis of the results. The research results show that: (1) The accident rate is relatively high at 7, 10 a.m., 5 to 6 p.m. and 9 p.m. (2) Collisions at certain angles and side collisions occur significantly more frequently than other types. (3) Compared with other types of roads, rural self-built roads have a significantly higher probability of accidents. (4) Road conditions with visibility between 100 and 200 meters are more likely to cause traffic accidents. (5) Unmarked mixed motor vehicle and non-motor vehicle lanes are high-risk areas for accidents. The results of the study can provide a reference and basis for the development of prevention and control strategies for rural road traffic accidents in Guangxi, and provide a certain reference value for rural road safety in China.