摘要
为了准确掌握未来交通流量的变化趋势,提高高速公路路网的管理效率,采用经验模态分解(EMD)和自回归滑动平均(ARMA)模型,提出了一种短时交通流量预测方法。根据高速公路收费站数据,使用EMD将统计的时间序列分解为有限个固有模态分量,对固有模态分量使用模糊C均值聚类,再采用ARMA将聚类后的固有模态分量进行预测,最后把每个分量预测值求和得到交通流量预测值。实例仿真计算表明,该算法比直接使用ARMA模型进行预测具有更高的预测精度,是一种有效的短时交通量预测方法。
In order to accurately grasp the trend of future traffic flow, improve management efficiency of the highway network, an approach to short-term traffic flow prediction is proposed based on Empirical Mode Decomposition （EMD） and auto regressive moving average.（ARMA） model. According to highway toll station data, the statistical time series is decomposed into different Intrinsic modes by EMD, then the intrinsic mode components are clustered by fuzzy clustering. These different clustered components are predicted by ARMA model. Finally, the traffic flow is obtained by summation the predictive value of each component. Experiment of simulation show that the proposed algorithm has higher predictive accuracy than the direct use of ARMA model, and it is an effective short-term forecasting method.
出处
《公路》
北大核心
2015年第5期124-129,共6页
Highway
基金
四川省交通科技项目
项目编号2013c7-1
