A HYBRID DEMAND FORECASTING MODEL BASED ON EMPIRICAL MODE DECOMPOSITION AND NEURAL NETWORK IN TFT-LCD INDUSTRY

Abstract

Demand forecasting plays an important role in the thin-film transistor liquid crystal display (TFT-LCD) industry. A hybrid approach is proposed for demand forecasting by combining empirical mode decomposition (EMD) and neural networks. From the signal analysis point of view, demand can be considered as a nonlinear and nonstationary combination of different frequencies. Every demand can be represented by one or several frequencies. The process of the proposed approach first decomposes the historical demand data into a finite set of intrinsic mode functions (IMFs) and a residual through EMD. Then, these IMFs are input into a back-propagation neural network (BPN) and the corresponding demand is used to predict these IMFs. Finally, the demand is forecasted by summing the predicted IMFs. The results show that the proposed model outperforms the single BPN model without EMD preprocessing and the traditional autoregressive integrated moving average (ARIMA) models.