| Grey system model is a widely used prediction method, but is always misused in practical problems with low forecasting precision. This paper narrates the experience of the development of the grey system model, tries to describe the law, finds out the mistakes in it, adapts the conventional theory, and enlarges the application range of GM(1,1)in the end.This paper divides the development of the grey system model theory into three stages and deals with seven key questions. It shows the real reason that the conventional theory (GM(1,1)) has systematic errors by discussing the grey accumulation formation of the original data, the grey derivative's whiting value, and the initial value problem. Furthermore, this paper comes to conclude mat the application range of GM (1,1) is when the original data follow accurate exponential distribution and then proposed an optimal model building procedure. There are four improving methods discussed in the follow, which are data transformation, the selection of model style, grey-Markov model, and grey-GARCH model.There are four chapters in this paper. The first chapter consists of the paper's purpose, meaning, aim, the former research works and the introduction of general Grey system model. The other three chapters are the main sectors of the paper. Each chapter stands by a stage of the development of the grey system model theory. Every key question is dependently discussed in certain paragraph. Every paragraph has similar structure with pointing out a question firstly, then finding out the former research works, the author's views and works at lastIt may be four new ideas followed:1. It firstly sets up a research frame of grey accumulation formation of the original data, concentrates on the original data with non- exponential distribution and concludes that this kind of original data is out of the application range of GM(1,1).2. Influenced by grey-Markov model, it proposes a new mixed model: grey-GARCH model that used to predict the time series in finance. |
PDF Full Text Request