基于ARIMA-EGARCH模型的交通事故问题的研究

RESEARCH ON TRAFFIC ACCIDENT PROBLEM BASED ON ARIMA-EGARCH MODEL

为了更好地掌握交通事故的现状和发展趋势,减少交通事故带来的直接损失和人员伤亡,以1990-2017年的全国交通事故发生起数为研究对象,建立ARIMA(0,1,4)模型进行分析和预测.利用方差齐性检验得残差序列方差非齐性,针对该模型的异方差问题,建立了ARIMA-EGARCH(1,1)模型.最后,以2017年的数据做为考核样本来检验模型的精度,并对2018-2019年的数据进行预测.通过与ARIMA(0,1,4)模型的拟合效果对比可以看出:ARIMA-EGARCH模型的预测结果比ARIMA模型精度更高,误差更小,可以对未来交通事故数据进行更好的预测.

This paper analyze traffic accident incidents to reduce the direct losses and casualties.We establish ARIMA(0,1,4)model for analysis and prediction using data covering the period from 1990 to 2017.The results from the ARIMA models reveal the presence of heteroskedasticity.Recognizing the heteroscedasticity of the model,we estimate primitive sequence using a ARIMA-EGARCH(1,1)model.Finally,we inspect the forecasting performance of the model by the data of 2017,and predict the data of 2018-2019.The results show that the ARIMA-EGARCH model outperforms the ARIMA model in terms of more accurate and less error.Through this model,the future traffic accident data can be better predicted.