为了分析高速公路平纵线形和其他交通相关因素对交通安全的影响,从时间和空间角度选取9个自变量,构建基于负二项回归的高速公路交通事故时空分析模型。以某高速公路典型危险路段为例,进行模型拟合与检验。结果表明,基于负二项回归的高速公路交通事故时空分析模型拟合效果良好,事故率与路段弯曲比和连续下坡累积效应呈正相关,与平均坡度和大车比例呈负相关。
We constructed a negative binomial regression based expressway traffic accidents temporal-spatial analysis model from nine independent variables in time and space domains to analyze the impacts of expressway horizontal and vertical alignments and other traffic related factors on traffic safety. We then perform model fitting and test, taking a typical dangerous expressway section as an example. Results show that the negative binomial regression based expressway traffic accidents temporal-spatial analysis model has better fitting effect. Accident rate is positively relative to curve proportion and cumulative effect of continuous downhill, while negatively relative to mean gradient and large car orooortinn.