Research On Some Issues Of Grey Model

Grey system theory is a emerging information system discipline processing "small sample", "poor information" uncertain systems, it has been widely used in the social, economic, energy, military and other fields and showed excellent results.Based on theory analysis of the grey GM(1,1) model, grey Verhulst model and gray relational analysis theory, this paper proposed different improved methods for these three aspects in order to achieve better prediction analysis results, the main contents:1. Preprocessing of original sequence data. From the point of grey GM(1.1) basic theory, we proposed the cosine function transformation method in view of the original data through standardized processing, it is demonstrated from mathematical method that the preliminary standardized processing and cosine function transformation can improve the original data sequence smoothness, thus improve the model prediction accuracy.2. From modeling mechanism, we adopted combination optimization method. Through fusion of exponential smoothing, background value optimization, initial value correction and metabolism model, the prediction ability of model was effectively improved.3. According to grey Verhulst model error sources, this paper refactor the grey derivative of model using method of undetermined coefficients, and unbiased grey Verhulst model is set up using modified grey derivative.4. In view of negative correlation between the proposed objects, a new gray-related degree was proposed, it can improve the effectiveness of gray relational analysis both in time and space.