Construction of Improved GM(1,1) Model and Its Use of Peanut Production and Consumption

Abstract

This paper aims at the influence of the disorder original data on the prediction accuracy of the GM(1,1) model, a waveform data generation method is proposed to make the disorder data become monotonically increasing order data. Then the initial condition of the time response function is reset according to the new information first principle, and the accuracy of the improved model is tested. The prediction accuracy shows that in terms of peanut yield, the posterior error ratio and average relative error obtained by using the improved GM(1,1) model are 0.117 and 0.0499, respectively, which are smaller than the posterior error ratio and average relative error of traditional models. In terms of consumption, the posterior error ratio predicted by using the improved GM(1,1) model is 0.078, which is smaller than the posterior error ratio of traditional models, which is 0.093. Compared to the traditional model, the improved GM(1,1) model can better predict China’s peanut production and consumption with higher prediction accuracy. This article uses an improved GM(1,1) model with high prediction accuracy, and China’s peanut production and consumption will continue to increase from 2023 to 2025.