Research On The Application Of Combined Gray Model

The study of social and economic system needs the support of forecasting theory. We can create the quantitaitive model, which reflects the law of system development, through the analysis of historical data, so that we can supply the powerful tool to analyze system characters and developing trend.Gray model is a typical model of tendential and analytical. It is suitable to those forecasts in social economic system, which has much factors, complex structure, large extent and high layers. The gray model also has particular efficiency to bate the unregulariy of date and discover the development regulation of system. If we join the gray model with other models to assay and predict, the ability of forecast can be amplified and the precision of prediction can be improved. In the past several years, many savants think much of gray combined model and some outcomes are achieved, which shows that the gray combined model is a valuable problem. But the research of gray combined model is still not perfect, it is necessary to research, which is meaningful to the developmemt of forecasting theory.This paper presents the following innovation:1. This paper proposes the gray-index moving combined forecast model, which combines the single-index moving method with gray disaster prediction theory. The joint model solves the problem of deviation and lag in single- index moving forecasting method in some degree, and prediction accuracy and reliability is improved.2. In the fourth chapte of this paper, the coupling forecast model of gray-multiple regressive analysis is founded. The independent variables of multiple regressive model are decided by gray correlation analysis, which make multiple regressive model more unfeignedly reflect the truth of future and prediction accuracy improved. On this base, the optimal multiple regressive model is got. Then, the gray prediction method is infiltrated into the operational process of optimal multiple regressive model and the advantage of each method is joint.3. This paper also presents the gray-artificial neural network combined model, which combines the gray metabolic model with the modified artificial neural network, and the weight coefficient of combined forecasting method is decided by least squares method. The combined method is applied to forecast the electricity consumption in Jiangsu, results are presented, and prediction accuracy of the combined model is more...