论文部分内容阅读
就有效预防交通事故、提高道路交通效率而言,借助高精度的道路交通事故预测模型,准确分析事故原因是重要的基础性工作。首先基于偏相关分析方法,对影响事故起数、死亡人数和受伤人数这3个事故指标的11个因素进行相关性分析,确定最相关的影响因素及其线性相关性;然后利用偏最小二乘回归方法,对事故指标与影响因素之间的线性关系进行建模;进而基于非线性偏最小二乘回归方法,建立两者之间的非线性关系模型。通过对回归模型的精度分析,用偏最小二乘回归方法仅能对事故指标与影响因素之间线性关系准确建模,测定系数最大为0.98,相对误差最大为21.77%。用非线性偏最小二乘回归方法,对事故指标与影响因素之间的线性和非线性关系均能准确建模,测定系数最大为1.相对误差最大为4.23%。
In order to effectively prevent traffic accidents and improve the efficiency of road traffic, it is an important basic work to accurately analyze the causes of accidents with the help of high-precision road traffic accident prediction models. Firstly, based on the partial correlation analysis, 11 factors that affect the accident indicators, the number of deaths and the number of injuries were analyzed and the most relevant influencing factors and their linear correlations were determined. Then, partial least squares Regression method to model the linear relationship between the accident indicators and the influencing factors. Then based on the nonlinear partial least-squares regression method, a nonlinear relationship model between the two is established. Through the precision analysis of the regression model, partial least squares regression method can only accurately model the linear relationship between the accident index and the influencing factors. The maximum coefficient of determination is 0.98 and the maximum relative error is 21.77%. Non-linear partial least-squares regression method can accurately model the linear and nonlinear relationship between the accident index and the influencing factors, the maximum coefficient of measurement is 1. The relative error is 4.23%.