基于NSGM-ARIMA变权组合模型的粮食产量预测——以1983-2022年的湖南省统计年鉴数据为例  

Grain Yield Prediction Based on NSGM-ARIMA Variable Weight Combination Model:Taking the Statistical Yearbook Data of Hunan Province from 1983 to 2022 as an Example

作　　者：苑慧芳 赵学超 YUAN Huifang;Zhao Xuechao(School of Mathematics and Statistics,Zaozhuang University,Zaozhaung 277160,China)

机构地区：[1]枣庄学院数学与统计学院,山东枣庄277160

出　　处：《枣庄学院学报》2025年第2期25-36,共12页Journal of Zaozhuang University

基　　金：枣庄学院本科教学改革研究面上项目(NO.YJG23051,NO.YJG22054);教育部产学研协同育人项目(241001065175204)。

摘　　要：粮食产量预测是保障国家粮食安全和为稳定市场供应的重要依据,且播种面积是影响产量的重要因素。以湖南省粮食产量数据为例,首先,建立粮食产量与播种面积之间的新结构多变量灰色预测模型(new structure multi-variable grey prediction model,NSGM)和差分自回归移动平均预测模型(autoregressive integrated moving average model,ARIMA),并对模型进行分析;其次,在残差变权法的基础上,基于NSGM-ARIMA模型提出一种新的变权组合预测方法用于预测;最后,将新的变权方法与单个模型、等权组合法和残差变权法进行比较说明所提变权方法的有效性,可作为粮食产量短期预测的有效工具。The prediction of grain yield is an important basis for ensuring national food security and stabilizing market supply,and the sown area is an factor affecting yield.Taking the grain yield data of Hunan Province as an example,the new structure multi-variable grey prediction model(NSGM)and the autoregressive integrated moving average prediction model(ARIMA)are established between grain yield and sown area,and the models are analyzed;Based on the residual variable weight method,a new variable weight combination prediction method based on the NSGM-ARIMA model is proposed for prediction;The effectiveness of the proposed variable weight method is demonstrated by comparing it with a single model,equal weight combination method,residual variable weight method,and third-order average combination method,which can serve as an effective tool for short-term prediction of grain yield.

关 键 词：NSGM模型 ARIMA模型 残差变权法 变权组合预测法 

分 类 号：S114[农业科学—农业基础科学] O211.67[理学—概率论与数理统计]