Empirical Analysis And Forecast Of Xinjiang GDP Based On Optimized Grey Model

The GM(1,1)prediction model is the most widely used dynamic model in gray system theory,but sometimes it has disadvantages such as low prediction accuracy and stability.Therefore,some scholars optimize the traditional GM(1,1)model from different angles,but most of them only consider the role Single-direction optimization methods such as volume,background value,initial value,and so on,combine multiple aspects to comprehensively improve less.So this article starts from this,and explores multiple aspects of comprehensive improvement methods to further improve its prediction accuracy.This article briefly introduces the modeling method of the traditional GM(1,1)model,an d analyzes the shortcomings of the traditional model in the selection of the initial value and ba ckground value.Through the analysis of several types of improvement methods of the backgr ound value and the initial value,it finally gives Three models of single-direction and combine d optimization are used,namely:the background value is reconstructed by the Gauss-Chebysh ev formula,the initial value is dynamically selected by using the one-time cumulative value x¹(k)and the minimum square error of the actual measured value,and one a combination op timization method that further reduces the prediction error of the model by combining the two.Finally,the 2010-2019 GDP data of Xinjiang is used for prediction analysis and residual testi ng.The results show that the combined optimization model has high stability and accuracy,an d then use the model to scientifically predict Xinjiang's GDP level in the next three years.This article will be explained in five chapters.The first chapter introduces the research background and current situation of this article,and explains the process of continuous optimization and promotion of the GM(1,1)model;the second chapter mainly introduces the basic knowledge of gray theory and the modeling process of traditional GM(1,1)model The third chapter introduces the background value,the initial value and the method of combination optimization;the fourth chapter uses the Xinjiang data in the past ten years to verify the feasibility of each optimization method,and selects the most accurate model to predict the next three years of Xinjiang's GDP level;the last chapter is a summary of this article and puts forward the direction of future research.