Properties of NGM (1,1,v) and Its Optimized Model with Multiple Transformation

摘要

Studying parameters properties of grey prediction models is beneficial to simples the modeling processes of grey models and improve their modeling efficiency. On the basis of modeling mechanism of grey verhulst model, this paper redefines novel grey verhulst model (Abbreviated as NGM (1,1,v) model,) improves this grey model by putting forward an algorithm for optimizing the initial value of NGM (1,1,v)model, and builds an optimized NGM (1, 1, v) model (Abbreviated as ONGM (1,1,v)) accordingly. The changing rule of the modeling accuracy of two novel grey verhulst models as multiple transformations acting on sequence of system characteristic and sequence of the related factor affecting system is analyzed. Research results show that the modeling accuracies of NGM (1, 1, v) and ONGM (1,1,v) model are all in relation to multiple transformation of sequence of related factors affecting system, but they have nothing to do with the multiple transformation of sequence of systems characteristics. Research results conclude that multiple transformation acting only on characteristic sequences of systems can simplify, the complexity of constructing two novel grey verhulst models with unchanging their simulative accuracy and predictive accuracy.