Logistics Demand Prediction Based on ARMA Model Improved by Blind Number Approach

Firstly, this article introduces the concept of the blind number after presented ARMA model; the possibility of using blind number to improve a model was discussed. Secondly the measures and steps of ARMA model improved by blind number were given. Finally, the ARMA model improved by blind number approach with 1984–2003 data was built, and China's logistics capacity from 2004 to 2008 is predicted. The absolute errors between predicted values and actual values are 5.2%, 4.8%, 4.2% and 4.4%, superior to the prediction results of ARMA models. The results indicate that ARMA model improved by blind number has features of more comprehensive consideration on influencing factors, more macroscopic description on historical data and more accurate prediction on results. Therefore, utilizing this method on logistics demand forecasting is accessible and feasible.