Research Of Rail Transport Forecasting Model Based On Support Vector Machine

Railway traffic forecast is basic work of economic evaluation of railway construction project, and it is also the foundation to develop future railway transport development strategy. Accuracy of railway traffic forecast have a direct impact on project investment and efficiency, The rational allocation of transportation resources, Full utilization of transport capacity and so on. Therefore, there is of great significance to forecast railway transportation efficiently and accurately. Now, there are kinds of methods and models on railway traffic forecast at home and abroad, their prediction performance and applicable conditions are different from each other. Because of the imbalance in social and economic development of our country, regional level of development of rail transport systems in different regions are Inconsistent, when making rail traffic forecasts, we need to look at the coordinated development of railway transport and regional economic system, and select reasonable model.In this paper, two aspects of research work is done:On the one hand, this paper puts forward the coupling degree evaluation model of railway transport and regional economic system with reference to the coupling degrees function in physics. As the main mode of transportation to promote economic development, railway transport is the infrastructure and a strong driving force to support for the economic development of our country. There is a very important link between railway transport and regional economic. By coordinated development, they achieve the purpose of mutual promotion and common development. This paper puts forward the coupling degree evaluation model of railway transport and regional economic system, it provides a way for understanding the coordinated development of railway transport and regional economic system more comprehensively and accurately. And it also provides a reference for choosing the rail transport forecasting model reasonably.On the other hand, this paper introduces support vector machine (SVM) into the rail transport forecasting model, and establishes the rail transport forecasting model based on SVM multiple regression and SVM auto regression. SVM is a kind of machine learning methods based on statistical learning theory (SLT), can solve nonlinear, high dimensional, poor information and other issues. This paper establishes the rail transport forecasting model based on SVM, and uses of the relevant data of "Statistical Yearbook of Chengdu Railway Bureau" for simulation experiment. By comparison with linear regression model, BP neural network model, and gray system model, the results show that SVM is a kind of forecasting model with high accuracy, good stability, can be used for rail transport forecasting effectively.