基于ARMA平稳时间序列的道路交通事故预测

Prediction of road traffic accidents based on ARMA stationary time series

为更好地进行道路交通安全评价、道路交通规划以及事故防治,采用时间序列分析法,分析事故发生起数时间序列上的趋势性规律,进行数据预处理和模型的识别与检验,并建立ARMA(2,1)预测模型,运用2003年至2015年全国道路交通事故起数进行分析和建模,并对2011年至2015年数据进行预测.研究结果表明:预测的趋势与实际的趋势基本一致,最大误差率为3.905%,综合误差率仅为1.366%,精度较高,效果较理想.

In order to carry out road traffic safety evaluation, road traffic planning and accident prevention, the time series analysis method is adopted to analyze the trend rule in the time series of the occurrence of accidents. In addition, data preprocessing and model identification and testing were carried out. Then, ARMA （2,1） prediction model was established. The number of road traffic accidents in China from 2003 to 2015 was analyzed and modeled to predict the data from 2011 to 2015. The results show that the predicted trend is consistent with the actual trend. The maximum error rate is 3.905%, and the comprehensive error rate is only 1.366%. The accuracy is high and the prediction effect is satisfactory. The time series analysis and prediction method highlights the role of time factors in the prediction.