摘要
本文基于2000~2017年山西省粮食产量数据分别构建了灰色预测模型、残差修正的灰色预测模型以及灰色线性组合模型,并预留2018~2021年的粮食产量数据进行预测精度检验。研究表明:灰色线性组合模型优于单一的灰色预测模型。灰色预测模型、残差修正的灰色预测模型以及灰色线性组合模型的平均绝对百分比预测分别为12.01%、11.99%、4.49%,拟合精度分别为87.99%、88.01%、95.51%,灰色线性组合模型的预测精度较灰色预测模型其预测精度提高了62.61%。
This paper constructs a gray forecasting model, a residual-corrected gray forecasting model and a gray linear combination model based on 2000~2017 grain yield data in Shanxi Province, respectively, and sets aside 2018~2021 grain yield data for testing the forecasting accuracy. The study shows that the gray linear combination model is better than the single gray forecasting model. The average absolute percentage forecasts of the gray forecasting model, the residual-corrected gray forecasting model, and the gray linear combination model were 12.01%, 11.99%, and 4.49%, respectively, and the fitting accuracy was 87.99%, 88.01%, and 95.51%, respectively, and the forecasting accuracy of the gray linear combination model improved by 62.61% compared with that of the gray forecasting model.
出处
《统计学与应用》
2022年第6期1421-1430,共10页
Statistical and Application
