Logistics amount forecasting based on combined ARIMA and ANN model

The logistics amount of some enterprises has a dual characters of growth and seasonal fluctuation. Multiple seasonal ARIMA model has linear fitting ability and ANN has the ability of nonlinear relationship mapping. A combined forecasting model based on multiple seasonal ARIMA model and ANN model was proposed to overcome the defects of single model, and the prediction result shows that the combined forecasting model is superior to the single model in many performance aspects. Combined forecasting model offers a new effective method of logistics amount prediction.