Application of IGM(1,1)-TFN model in the forecasting of returns
IGM(1,1)-TFN 模型在产品回收预测的应用

To overcome the disadvantages of the current forecasting returns models,such as narrow application extent and requirement for more data,an improved GM(1,1)-transfer function-noise model is proposed in this paper.For the lack of return data in the first stage,the GM(1,1)model improving the generation method of neighborhood system is adopted,and forecasting is carried out by only using a few data.When data amounts reached to 50,the hidden quantitative relationship between sale quantity and return quantity is established by performing the transfer function-noise model.Its parameters are determined through model identification,parameters estimation and diagnostic checking.Then,forecasting return is accomplished.Finally,an example is provided to verify the rationalities and efficiencies of the model.