基于多元时间序列的交通流预测模型
|
摘要
交通流预测是智能交通系统的一项重要研究内容。本文考虑路网的复杂性,使用多元时间序列模型建模,针对贵阳市省医路口2016年4月上下游交通流数据:首先建立ARIMA模型,采用Ljung-Box方法对模型效果进行检验,模型未能通过显著性检验;然后根据车辆的基本通行规律,构建多元时间序列ARIMAX模型,第一步利用主成分回归建立上下游交通流回归模型,第二步对模型残差进行ARIMA建模,得到的ARIMAX模型能够通过LB检验;最后利用预测误差绝对值均值和预测误差百分比绝对值均值进行模型性能评价,构建的ARIMAX模型均优于ARIMA模型。
Traffic flow forecasting is an important research content of intelligent transportation system. This paper considers the complexity of road network,and uses multivariate time series model,aiming at the traffic upstream and downstream data of Shengyi intersection in April 2016 in Guiyang: above all the ARIMA model was established to fit the traffic flow data,the Ljung-Box method was used to test the model effect,the model could not pass the significance test; then according to the basic rules of traffic,the multivariate time series ARIMAX model was created,firstly this paper built a upstream and downstream traffic flow regression model by principal component regression,and next the model residuals are fitted by ARIMA,and the model is verified by LB test; finally,the ARIMAX model is better than the ARIMA model by using the mean absolute error and the mean absolute percentage error.
Traffic flow forecasting is an important research content of intelligent transportation system. This paper considers the complexity of road network,and uses multivariate time series model,aiming at the traffic upstream and downstream data of Shengyi intersection in April 2016 in Guiyang: above all the ARIMA model was established to fit the traffic flow data,the Ljung-Box method was used to test the model effect,the model could not pass the significance test; then according to the basic rules of traffic,the multivariate time series ARIMAX model was created,firstly this paper built a upstream and downstream traffic flow regression model by principal component regression,and next the model residuals are fitted by ARIMA,and the model is verified by LB test; finally,the ARIMAX model is better than the ARIMA model by using the mean absolute error and the mean absolute percentage error.
引文
[1]Ahmaed Mohamed S,Cook Allen R.Analysis offreeway traffic time-series data by using Box-Jenkins technique[J].Transportation Research Record,1979,722:1-9.
[2]Williams B,Durvasula P,Brown D.Urban Freeway Traffic Flow Prediction:Application of Seasonal Autoregres-sive Integrated Moving Average and Exponential Smoothing Models[J].TransportationResearch Record,1998,1644(1):132-141.
[3]臧利林,贾磊,杨立才,等.交通流实时预测的混沌时间序列模型[J].中国公路学报,2007,20(6):95-99.
[4]姚智胜.基于实时数据的道路网短时交通流预测理论与方法研究[D].北京:北京交通大学,2007.
[5]杨元元.基于时间序列模型的短时交通流预测的研究与应用[D].西安:西安电子科技大学,2014.
[6]罗媛媛.基于EVIEWS的短时交通流分析及预测[D].成都:西南交通大学,2009.
[7]葛志鹏,李锐,张健,等.基于时间序列与GSVMR模型的短时交通量组合预测[J].长安大学学报(自然科学版),2015(S1):222-225,234.
[8]邴其春,杨兆升,周熙阳,等.基于向量误差修正模型的短时交通参数预测[J].吉林大学学报(工学版),2015,45(4):1076-1081.
[9]汪宏晶,林曦晨,汤洪秀,等.多元时间序列模型及其应用[C]//2011年中国卫生统计学年会会议论文集.西安:中国卫生统计,中国卫生信息学会统计理论与方法专业委员会、中华预防医学会卫生统计专业委员会,2011(4):238-241.
[10]汪远征,徐雅静.多元平稳时间序列ARIMAX模型的应用[J].统计与决策,2007(18):132-135.
[11]李春燕,陈峻.基于ARIMAX模型的交通事故宏观预测[J].道路交通与安全,2009(01):18-22.
[12]程宇,夏若雯.汇率改革对人民币兑美元汇率中间价的影响——基于ARIMAX模型的实证分析[J].当代经济,2016(17):22-24.
[13]王燕.应用时间序列分析[M].北京:中国人民大学出版社,2012.
[2]Williams B,Durvasula P,Brown D.Urban Freeway Traffic Flow Prediction:Application of Seasonal Autoregres-sive Integrated Moving Average and Exponential Smoothing Models[J].TransportationResearch Record,1998,1644(1):132-141.
[3]臧利林,贾磊,杨立才,等.交通流实时预测的混沌时间序列模型[J].中国公路学报,2007,20(6):95-99.
[4]姚智胜.基于实时数据的道路网短时交通流预测理论与方法研究[D].北京:北京交通大学,2007.
[5]杨元元.基于时间序列模型的短时交通流预测的研究与应用[D].西安:西安电子科技大学,2014.
[6]罗媛媛.基于EVIEWS的短时交通流分析及预测[D].成都:西南交通大学,2009.
[7]葛志鹏,李锐,张健,等.基于时间序列与GSVMR模型的短时交通量组合预测[J].长安大学学报(自然科学版),2015(S1):222-225,234.
[8]邴其春,杨兆升,周熙阳,等.基于向量误差修正模型的短时交通参数预测[J].吉林大学学报(工学版),2015,45(4):1076-1081.
[9]汪宏晶,林曦晨,汤洪秀,等.多元时间序列模型及其应用[C]//2011年中国卫生统计学年会会议论文集.西安:中国卫生统计,中国卫生信息学会统计理论与方法专业委员会、中华预防医学会卫生统计专业委员会,2011(4):238-241.
[10]汪远征,徐雅静.多元平稳时间序列ARIMAX模型的应用[J].统计与决策,2007(18):132-135.
[11]李春燕,陈峻.基于ARIMAX模型的交通事故宏观预测[J].道路交通与安全,2009(01):18-22.
[12]程宇,夏若雯.汇率改革对人民币兑美元汇率中间价的影响——基于ARIMAX模型的实证分析[J].当代经济,2016(17):22-24.
[13]王燕.应用时间序列分析[M].北京:中国人民大学出版社,2012.