论文部分内容阅读
交通事故的发生不仅导致交通延误,而且会造成不同程度的人员伤亡及财产损失,因此预测交通事故的处置时间及其严重程度具有重要意义。本文抽取了某高速公路交通事件管理系统中记录的近三年的交通事故信息并进行数字化处理;采用方差分析的方法,深入分析高速公路交通事故处置时间和严重程度的关联因素及作用;对于事故处置时间的预测,采用多元逐步回归建模的方法找出主要关联因素并进行多元回归预测;对于事故严重程度的预测,采用定序logistic逐步回归方法构建三等级的事故严重程度预测模型。经样本检验,两类模型能够很好地预测交通事故处置时间及其严重程度。
Traffic accidents not only lead to traffic delays, but also cause varying degrees of casualties and property losses, so it is of great significance to predict the timing and severity of traffic accidents. In this paper, traffic accident information of a highway accident event management system is recorded and digitized in the past three years. The variance analysis method is used to analyze in depth the related factors and effects of the highway traffic accident disposal time and severity. The method of multivariate stepwise regression modeling found the main correlative factors and carried out multiple regression prediction. For the prediction of accident severity, the logistic stepwise regression method was used to build the three-level prediction model of accident severity. After the sample test, two kinds of models can predict the traffic accident disposal time and its severity well.