摘要
交通流预测是智能交通系统的基础,由于交通流量增大,造成交通堵塞。预测某段单位时间内的交通流,传统方法很难准确表示交通流量的时变性、突发性和非线性等变化规律,预测精度较低。为了提高交通流的预测精度,提出了多尺度小波支持向量机预测模型,并将之应用于交通流预测中。利用小波多分辨率分析,构造出多尺度小波核函数,实现了小波技术与支持向量机方法的结合。实验结果表明,支持向量机预测效果比神经网络要好,多尺度小波核函数比径向基核函数更优,多尺度小波支持向量机在交通流预测中具有应用的可行性。
Traffic flow forecasting is the basis of the Intelligent Transportation Systems, accurate traffic flow fore- casting plays a vital role in the development of intelligent traffic and transport information system. The traditional fore- casting method can not accurately describe the characteristics of time-varying, sudden and nonlinearity of traffic flow, thus traffic flow prediction accuracy is low. In order to improve the prediction accuracy of traffic flow, the paper pro- posed a prediction model based on the multiscale wavelet support vector machine, and applied to the traffic flow fore- casts. Based on wavelet multi-resolution analysis, a scaling kernel function with multi-resolution characteristics was constructed to implement the combination of the wavelet technique with support vector regression. The experimental results demonstrate that the predicted effect with support vector machine is better than the neural network and the pro- posed approach with multiscale wavelet kernel can provide more optimal performance than that with radial basis func- tion kernel, and the feasibility of applying MW-SVR in traffic flow prediction.
出处
《计算机仿真》
CSCD
北大核心
2013年第11期156-159,共4页
Computer Simulation
基金
江苏软科学计划项目(BR2012043)
关键词
智能交通
支持向量机
多尺度小波
交通流预测
核函数
Intelligent transportation
Support vector machine （ SVM ）
Multiscale wavelet
Traffic flow prediction
Kernel function
