Improvement Of Grey Forecasting Model And Its Application

Grey forecasting model is an important part of grey system theory, also is a widely used prediction method in the prediction theorey and application. Therefore, research on grey forecasting model is of great significance.At present, the problems of grey forecasting model are as follows: the improved method is complex, the scope of application is confined, and forecasting precision is not high. Against these problems, on the basis of summarizing the existing outcome of improved grey forecasting model, the paper study on improved grey forecasting model using function transformation and inverse transformation theory, grey system modeling theory and genetic algorithm theory. Particularly, the research on improving grey model's forecasting precision by increasing the smooth degree of modeling data series is given. And viability of improved methods is proved and effectiveness of improved methods is tested by practical examples. The primary research content and results are obtained as followes:1. The smooth degree of modeling data series is one of the important factors that impact forecasting precision of grey model. Thus, on the basis of translation transformation and stretching transformation, the paper presented a linear function transformation improved grey forecasting model with parameters and proved that the improved method can increase the smooth degree of modeling data series. Finally, numerical example is given and fitted effect and forecasting precision of improved model is analysised.2. In the paper, on the basis of logarithm function transformation and linear function transformation can increase the smooth degree of modeling data series, logarithm function-linear function transformation improved grey forecasting model with parameters and linear function-logarithm function transformation improved grey forecasting model with parameters are given. It has been proved that the two transformation methods can increase the smooth degree of modeling data series. The calculation results of practical example show effectiveness of the two improved methods.3. Considering that some power function transformation and linear function transformation can improve forecasting presicion of grey model by increasing the smooth degree of modeling data series, the paper presented that power function-linear function transformation improved grey forecasting model with parameters and linear function-power function transformation improved grey forecasting model with parameters. The proof that the two transformations can increase the smooth degree of modeling data series is given. Finally, practical example's calculation result show that the two improved methods is effective.4. Grey forecsating model improved by function transformation can incease precision of short-term prediction and new information and equal dimensional model with suitable dimension can increase presicion of long-term prediction. The paper presented that new information and equal dimensional grey model based on power function-linear function transformation with parematers. Finally, practical example shows that the improved model is feasible.