Abstract. The purpose of this paper is to further improve accuracy of the grain yield prediction and enhance the robustness of the prediction algorithm. The method mainly involves the knowledge included Gray Theory and Multiple Linear Regression. Readers can refer to Headline 2 to understand. Firstly, this method analyzed the factors that affected the grain yield, smoothed the data of the previous influence factors exponentially, used the gray theory to iteratively predict the new impact factor data, and then automatically selected the factors with high correlation degree through the influence factor correlation analysis. Finally, different treatment processed grain yield data to reduce the fluctuation of the data. To predict future grain yield used multiple regression and residual correction. You can check the title 3 for a detailed solution. The forecasting method proposed in this paper has a good prediction effect, and the relative error of annual average prediction is less than 5%.
In order to predict the grain yield of the country accurately, considering the periodical fluctuation of the data, the method of time series is used. Firstly, the stability and the relativity of the yield series from year 1980 to 2009 are analyzed, and the first-order difference of which is calculated to get a stationary series. Then, after comparing the value of AIC of different models, the forecasting model ARIMA(5,1,5) is selected as the best one, and the performance of which is tested. Lastly, the grain yields from year 2010 to 2012 are predicted by three different methods, the results shown that, the prediction error of the model ARIMA(5,1,5) is 4.478%, the error of the grey model GM(1,1) is 6.78%, and the error of the second exponential smoothing method is 7.682%, thus, the model ARIMA(5,1,5) is more suitable to forecast the grain yield in short-term.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.