短时交通量预测是智能交通系统提供交通信息、诱导交通与制定控制策略的重要基础. 基于小波分析与支持向量回归机( SVR)预测,提出一种基于小波-SVR模型的高速公路短时交通量预测方法. 该方法采用小波分解与重构算法,将交通量原始信号分解为逼近信号和细节合成信号,利用SVR对2种信号分别进行建模分析,最后合成预测结果. 应用该方法可实现时间间隔为5 min的交通量预测. 实例分析表明:与直接应用SVR模型相比较,小波-SVR模型各项评价指标更优,其为交通量实时准确预测提供了更为科学的方法.%Forecast of short-time traffic volume is important basis for intelligent traffic system to provide traffic information, guide traffic and draw up control policies. Based on wavelet analysis and forecast of support vector regression ( SVR) , this paper proposes a method for forecasting short-time traffic volume on expressway based on wavelet - SVR model. This method adopts wavelet decomposition and reconstruction algorithm to decompose the original signals of traffic volume into approximation signal and detail synthesized signal, conducts modeling analysis of two signals by means of SVR, and finally synthesizes forecast results. This method can realize forecast of traffic volume at an interval of 5min. analysis of examples shows that in comparison with direct application of SVR model, all evaluation indices of wavelet-SVR model are more superior, and provide more scientific methods for real-time and accurate forecast of traffic volume.
展开▼
