An Improved Grey Forecasting Model with Three Parameters and Its Application

The grey forecasting model with three parameters (GFM_TP) is an important and useful grey model which can achieve unbiased simulation and prediction for both homogenous or non-homogenous exponential sequences and linear function ones. However, its precision is poor when an original sequence for modeling has the feature of random oscillation. Actually, the smoother the sequence is, the higher the simulation precision is. To this end, this paper applies an existing smoothing algorithm which can compress the amplitude of random oscillation sequence to improve the smoothness of a raw modeling sequence. Then an improved grey forecasting model with three parameters (IGFM_TP) is deduced. Finally, we employ the new model to forecast the electricity demand of a city in western China, and compare the simulation values and errors with those of GFM_TP, GM(1, 1) and DGM(1,1), the results shows the new model has the best simulation effect. Research findings in this paper have an important significance to improve the simulation and prediction performances of grey prediction model.