基于Bayes-ARIMA的景区公路短时交通流量预测
Prediction of Short-term Traffic Flow of Scenic Roads Based on Bayes-ARIMA
王代君 1李明 2鹿守山2
作者信息
- 1. 江西省交通科学研究院有限公司 南昌市 330000;长大桥梁建设关键技术及装备交通运输行业研发中心 南昌市 330000;江西省桥梁工程重点实验室 南昌市 330000
- 2. 江西省交通科学研究院有限公司 南昌市 330000
- 折叠
摘要
为方便景区公路交通组织及资源调度,提出了一种基于贝叶斯估计和ARIMA(Auto Regressive Integrated Moving Average model,差分自回归移动平均模型)的短时交通流量预测模型Bayes-ARIMA.通过ARIMA模型捕捉车流量时间序列特征,再通过贝叶斯方法引入其他时空因素的影响,充分利用2种模型的优势对车流量进行联合预测.结果表明:贝叶斯方法能够拟合交通流量的整体趋势,但在细节波动上的拟合精度明显不足,部分有用的细节信息丢失在残差序列中.ARIMA模型可以有效提取并还原贝叶斯预测残差序列中的有用信息,修正贝叶斯预测结果.与贝叶斯估计或ARIMA单独使用时相比,Bayes-ARIMA模型的均方根误差和绝对平均误差均有显著下降,表明Bayes-ARIMA组合模型的综合性能优于贝叶斯估计和ARIMA单一模型.
Abstract
In order to facilitate traffic organization and resource scheduling in scenic places,a short-term traffic flow prediction model Bayes-ARIMA is proposed based on Bayesian estimation and ARIMA(Auto Regressive Integrated Moving Average).The ARIMA model is used to capture the characteristics of the time series of traffic flow,and the Bayesian method is used to introduce the influence of other space-time factors,making full use of the advantages of the two models to jointly predict the traffic flow.The results show that Bayesian method can fit the overall trend of traffic flow,but the fitting accuracy of detail fluctuation is obviously insufficient,and some useful detail information is lost in the residual sequence.The ARIMA model can effectively extract and restore the useful information in the Bayesian prediction residual sequence and modify the Bayesian prediction results.Compared with Bayesian estimation or ARIMA model uses alone,the root mean square error and absolute mean error of Bayes-ARIMA model are significantly decreased,indicating that the comprehensive performance of Bayes-ARIMA combined model is better than Bayesian estimation or ARIMA single model.
关键词
智能交通/短时交通流量/预测模型/贝叶斯/ARIMAKey words
ITS/short term traffic flow/prediction model/Bayes/ARIMA引用本文复制引用
基金项目
江西省03专项及5G项目(20212ABC03W04)
江西省交通运输厅科技项目(2022X0043)
出版年
2024
NETL
NSTL
万方数据