A GROUP OF FORECASTING METHODS COMPARETIVE ANALYSIS

摘要

The forecast is the base of scientific decision-making, how to choose the right forecasting method, is very important for forecasting results. This paper uses the data of total freight in 1978-2006, and using the methods such as ARIMA, Gray forecasting and it's amend methods, as well as the BP neural network and gray support vector machines to forecast and determinate the error, and finally provides a simple comments for various methods. We proved the conclusion that the complicated methods are not necessarily better than the simple one once again. The paper has a certain theoretical and practical significance for the further study of combination forecasting methods and laid the foundation for further research.