Study On Methods Of Section Average Velocity Short-term Forecast Based On Support Vector Regression

Traffic control and traffic guidance are important parts in the Intelligent Transportation System ,those are effective means and methods to solve the traffic congestion, traffic accidents and other issues of traffic management. Traffic control and traffic guidance are basis of forecast accurately of the various traffic parametres. Traffic flow, traffic density and traffic velocity are all important traffic parametres . Therefore, the section average velocity short-term forecast, not only provides significant reference for real-time traffic control and traffic guidance,but also offers good traffic information services for travelers.This paper, based on the characteristics of floating car data, concludes the types of invalid data, gives an effective methods for data process, analyses the characteristics of kinds of support vector machine regression,the kernel function and the parameters of the model. Meanwhile, for the peculiarity of section average velocity forecast ,such as small samples, nonlinear, high-dimensional,the paper brings the methods of support vector regression into section vehicles velocity forecast,which is based on the floating car data ,and constructs section average velocity short-term forecast model. Additionally, certain section in HanZhou is taken as an example,using the LibSVM2.84 software package,to test and verify the practicability of the section average velocity short-term forecast model. Calculation results show that the method is feasible and effective.