In transportation engineering, freight volume forecast is the basis
for formulating relevant policies, preparing the transportation development programs and daily management. As an important
indicator to reflect the freight transportation demand, the prediction
research and analysis for freight volume has both practical and
theoretical significance. The forecast methods of the freight transportation demand can be divided into two categories: the qualitative forecast methods and the quantitative forecast methods. The qualitative forecast methods include specialistic forecast method, jury of executive opinion method, market investigation method and Delphi method. The specialistic forecast method is a commonly used method which is time-effective, simpler, and easier. Nevertheless, the qualitative forecast methods are less accurate and have one-sideness. On the other hand, the quantitative forecast methods include time series smoothness method, regression forecast method and grey forecasting method, etc. In addition, scholars introduce artificial neural network (ANN) and an intelligent fuzzy regression algorithm into the freight transportation forecast. Furthermore, partial least squares regression (PLS) method plays an important role in prediction. A multinomial probit model with spatially and temporally correlated error structure is proposed for freight demand analysis in tactical or operational planning applications.