Freight Volume Forecasting Based On SVM And Grey GM(1,1)

As a pillar industry in China,railway transportation plays a very important role in China's social and economic development.It is a transportation mode suitable for large-scale personnel flow and material transportation.In today's society,the logistics industry has developed rapidly,it brought tremendous pressure and competition to China's railway transportation.With the advancement of China's “the Belt and Road Initiatives”,it has further promoted the improvement of domestic industry and trade,and it made domestic and foreign trade more frequent,and brought development to China's waterway transportation.This makes the prediction of railway freight volume more meaningful.And it can not only provide certain support for the operation management department and the transportation enterprise to properly equip the transportation force,but also provide a reliable basis for the relevant planning and layout of the freight transportation system,thereby improving the economic and social benefits of freight transportation.In order to make the freight organization more efficient in arranging cargo transportation,it is necessary to make a relatively accurate prediction of freight volume.The main forecasting methods include time series forecasting,fractal theory,support vector machine and neural network,etc.In this paper,the particle swarm optimization algorithm is used to optimize the support vector machine to forecast the railway freight volume and the water freight volume.Support vector machine can better describe the nonlinear and random characteristics of freight volume when the data is relatively small,so that the prediction accuracy can be improved.Firstly,this paper gives a brief introduction to the theoretical basis for the main methods of support vector machine,particle swarm algorithm and grey forecasting,and it lays a foundation for the establishment of the subsequent forecasting model.Secondly,the data analysis for the railway freight volume?the railway freight turnover and the water freight volume of the research object is carried out.According to its instability,the difference calculation is performed,and the corresponding ARIMA model is fitted and used for prediction.The example shows that the model is more suitable for short-term forecasting.Then,the phase space reconstruction for the one-dimensional time series data is used as the input of the forecasting model.The cross-validation and particle swarm optimization are respectively used to optimize the parameters of the support vector machine and establish the corresponding forecasting model.Then the grey GM(1,1)model is used to make two kinds of combined forecasting.Finally,the same forecasts are made for the railway freight volume?the railway freight turnover and the water freight transport volume,and the prediction results ofvarious methods are compared and analyzed.It is proved that the method of combined prediction using a support vector machine and a grey GM(1,1)is effective.