A new forecasting method for time continuous model of dynamic system

摘要

Usually a linear differential equation is used to represent continuous dynamic systems, but a linear difference equation is used to represent discrete dynamic systems. AGO is one of the most important characteristics of grey theory, and its main purpose is to reduce the randomness of data. A linear differential equation, instead of a linear difference equation, is used to replace the grey differential equation to analyze discrete systems in this paper. Approximating a k-order derivative by operating after spline curve fitting of 1-AGO data, a model is directly established by means of the least square method. ARIMA models are used to analyze the leading indicator in advance, and the Fourier series with suitably chosen values of parameters is used in the fitting of leading indicator. This model is called the GDM(2, 2, 1) model.