Overview of Logistics Demand Forecasting Methods

Accurate forecasting of logistics demand is great theoretical significance and practical application value for the formulation national policies satisfaction actual in industry. From modeling form, existing methods are divided into four categories: single traditional method, intelligent combined method mixed method. Among them, mainly include simple time series regression analysis, mathematical statistical methods, etc.; involve gray neural network, support vector machine their improved forms; summarized three forms: linear combination results, nonlinear modified results; hybrid optimization algorithms with prediction data dimensionality reduction techniques mining methods. The major types reviewed, each model evaluated terms principles, advantages disadvantages, applicability, order to find suitable different tasks researchers.