Forecasting urban traffic flow by SVR

In recent years, the support vector regression model (SVR) has been widely used to solve nonlinear regression and time series problems. The paper proposes short-term traffic forecasting model based on support vector regression, the traffic volumes at preceding periods of time and upstream and downstream are considered as input, traffic volumes at current period of time are considered as output. The traffic volumes forecasted at several periods of time in the future are available by inputting the traffic volumes. Selection of kernel function is a pivotal factor which decides performance of SVR. The RBF kernel function is most widely used in SVR. There are two parameters in this function: the penalty parameter C and the kernel parameter Ȗ. This paper chooses a method to select the optimization parameters (C, Ȗ). A numerical example of traffic flow data from Qing Nian Da Jie, Peaceful Districst, Shenyang City, Liaoning Province, China is used to elucidate the forecasting performance of the proposed SVR model.