交通运输系统工程与信息

交通运输系统工程与信息 ›› 2013, Vol. 13 ›› Issue (2): 48-56.

• 智能交通系统与信息技术 • 上一篇    下一篇

基于OLAM的高速公路交通量多维预测研究

钱超1，许宏科*1，徐娜2，代亮1，程鸿亮1   

  1. 1. 长安大学 电子与控制工程学院，西安 710064； 2. 西安公路研究院，西安 710054
  • 收稿日期:2012-09-25 修回日期:2013-02-04 出版日期:2013-04-25 发布日期:2013-04-27
  • 作者简介:钱超（1984-），男，江苏新沂人，博士生.
  • 基金资助:

    国家自然科学基金项目（60804049）；教育部创新团队发展计划资助项目(IRT1050)；中央高校基本科研业务费专项资金资助项目（CHD2012JC056).

OLAM Based Multi dimensional Prediction of  Expressway Traffic Volume

QIAN Chao 1, XU Hong-ke 1, XU Na 2, DAI Liang 1, CHENG Hong-liang 1   

  1. 1.School of Electronic and Control Engineering, Chang’an University, Xi’an 710064, China;2.Xi’an Highway Institute, Xi’an 710054, China
  • Received:2012-09-25 Revised:2013-02-04 Online:2013-04-25 Published:2013-04-27

PDF

被引次数

9

摘要：

OLAM是联机分析处理与数据挖掘的有机结合，本文以高速公路收费数据为基础，提出一种基于OLAM实现高速公路交通量多维预测的方法.该方法构建了多维数据雪花模型，建立起收费数据的数据仓库并得到交通量多维统计结果；在构建季节ARIMA预测模型过程中，检测出因节假日、恶劣天气导致的交通量异常值并对模型进行修正；最后利用修正后的模型实现了交通量的预测.与一般季节ARIMA模型相比，修正后模型的白噪声方差和AIC值显著降低，数据拟合程度明显提高.实验结果表明，该方法具有较高的预测精度，其中MAE和MAPE分别为50.43和1.59%，能够满足高速公路管理部门利用收费数据分析、预测交通量时空变化趋势的要求，从而为制定各项政策提供理论依据和决策参考.

关键词: 公路运输, 多维预测, OLAM, 收费数据, 交通量, 季节ARIMA模型, 数据挖掘

Abstract:

The online analytical mining (OLAM) is the organic combination of online analytical processing and data mining. On the basis of expressway tolling data, this paper proposes a method of multidimensional prediction of expressway traffic volume based on the OLAM. The method formulates the snowflake schema of multidimensional data. It also establishes the data warehouse of tolling data and gets multidimensional statistics of traffic volume. In the seasonal ARIMA predicting model, traffic outliers caused by holidays and severe weather are detected and the predicting model is modified. Finally, the prediction of traffic volume is realized by the improved predicting model. Compared with the general seasonal ARIMA model, the white noise variance and AIC value of the model is significantly reduced and the fitting degree of data is obviously improved. The experimental results show that the proposed method provides high prediction accuracy and the MAE and MAPE are calculated to be 50.43 and 1.59%, respectively. This not only assists the expressway administrations to analyze and predict the spacetime changing trend of traffic but provides theoretical foundation and decision support for the work of making policies.

Key words: highway transportation, multidimensional prediction, online analytical mining (OLAM), tolling data, traffic volume, seasonal ARIMA model, data mining

中图分类号: