摘要
基于1989-2019年的社会消费品零售额数据,在建立多元线性回归模型、Holt-Winters非季节指数平滑和多项式拟合三种单项预测模型的基础上,将广义诱导因子引入,分别选取不同参数值的GIOWA算子,以最小误差平方和为最优性准则建立组合预测模型。结果显示,基于GIOWA算子的组合预测模型的预测效果高于三种单项预测模型。通过建立的组合预测模型预测了未来5年我国的社会消费品零售额。
The paper uses the data of the total retail sales of social consumer goods in China from 1989 to 2019,and establishes three single forecasting models:multiple linear regression model,Holt-Winters non-seasonal index smoothing and polynomial fitting model.In order to improve the prediction accuracy,the generalized induced factor is introduced.The paper establishes the combined forecasting models by selecting GIOWA operators with different parameter values and taking the minimum sum of squared errors as the optimality criterion.The results show that the prediction effect of the combined forecasting model based on GIOWA operator is better than that of the three single forecasting models.In the end,we predict the total retail sales of social consumer goods in China in the next five years.
作者
余凡
吴礼斌
Yu Fan;Wu Libin(Anhui University of Finance and Economics,Bengbu 233000,China)
出处
《廊坊师范学院学报(自然科学版)》
2020年第4期59-65,共7页
Journal of Langfang Normal University(Natural Science Edition)
关键词
GIOWA算子
组合预测
社会消费品零售额
GIOWA operator
combined forecasting
total retail sales of social consumer goods