In this paper, the regression forecast methods are used to forecast
the freight transportation demand based on the freight volume data
set, the gross domestic product (GDP) data set and the fixed
investment data set. Regression analysis methods are fit for short
term and mid-term forecast. They have been applied in many fields
such as analysis of data groups, estimation and statistics tests, analysis of relationships between influential factors and prediction target values, and the study of the accuracy of forecasting results.
In this paper, the fitted freight transportation model is constructed
based on regression analysis methods and then these methods help us to predict future trend of the freight volume. We make a comparison among the models established by using multiple linear regression (MLR), nonlinear regression (NLR) and simple linear regression (SLR) and then the best model is selected out, in order to improve the model accuracy. The methods are superior to the qualitative forecast methods in the facet of freight volume prediction. Based on the proposed methods, we can forecast the future value and error together with the accuracy of forecast result by analysing the mathematical relations among variables and the relationship between the influential factors and prediction target values. The models built based on regression analysis can be used in analyzing various freight transportation issues. It also offers suggestions for future research in freight transport demand. This paper can be generalized and extended to analyze the trends of freight transportation market in other parts of China and possibly elsewhere. The freight of Shanghai is critical to regional
economy. Based on the data they utilized, we choose GDP and
fixed investment as the influential factors and then build the models.