Analyses on Accuracy and Sensitivity of Grey Energy Model

Abstract

The grey theory carries out the system analysis and prediction on the basis of the accumulated data of original series. The accumulated generating operation (AGO) is one of the most important characteristics of grey theory. Its main purpose is to reduce the randomness of original data. Therefore, the grey theory is more suitable for analysis and prediction of the system whose regularity of the original series is poor and/or whose original information are lacking. The improved grey model by the equal-dimensional replenishment and residual modification with Markov-chain sign estimation further improve the accuracy of grey model. In this paper, we first simply introduce the mechanism of grey model and grey analysis and predictionmodel on China's energy system, then discuss the accuracy and variable sensitivity of this model. The results show that grey energy model has a higher simulating and forecasting accuracies. Along with GDP growth, the final energy consumption in China shows three stages: rapid growth stage, slow growth stage and slow declining stage. The sustaining period per stage is closely related with GDP growth rate. The longer the used original series is, the larger the future energy consumption is. The results of sensitivity analyses show that grey energy model is stable for GDP and the length of original series.