Neural Network Based Parameters Whiten Method Study of Gray Differential Equations

摘要

Building a relevant mathematical model for a certain system is a main aim for studies on a system. This paper applies neural network to whiten parameters in gray differential equations. After summarizing the main characters of these parameters, whiten procedures of GNNM(1,1) and GNNM(1,2) are deduced. Then the final gray differential equations can be expressed. An application shows the precision and efficiency difference between traditional method and method explained in this paper. And proves the priorities of this whiten method.