基于函数性数据分析的中国社会消费品零售总额数据的预测  

Prediction of Chinese total retail sales of consumer goods based on functional data analysis

作　　者：李小星 徐永利[1] LI XiaoXing;XU YongLi(Faculty of Science, Beijing University of Chemical Technology, Beijing 100029, China)

机构地区：[1]北京化工大学理学院,北京100029

出　　处：《北京化工大学学报（自然科学版）》2018年第3期107-112,共6页Journal of Beijing University of Chemical Technology(Natural Science Edition)

基　　金：国家自然科学基金(11571031);国家留学基金

摘　　要：基于数据的函数性特征,采用函数性数据分析方法对中国社会消费品零售总额进行预测。将总额数据分解为长期趋势成分和季节性波动成分分别进行预测。对长期趋势成分引入居民可支配收入数据来辅助预测,对于季节性成分采用自适应选择权重的加权预测法,将两项预测值的和作为总额预测结果。真实数据上的实验结果表明,本文提出的预测方法预测误差小,且预测结果具有很好的可解释性。The functional data analysis method is used to predict the data of Chinese total retail sales of consumer goods. The total data is decomposed into a long-term trend component and a seasonal fluctuations component,and then the predictions for these two components are carried out respectively. For the long-term trend component,the data of disposable personal income is introduced to support the forecast. For the seasonal component,a weighted selection method with adaptive selection weights is used. The sum of the two predictions is taken as the total forecast. The experimental results using the real data show that the proposed method can give smaller prediction errors and the predictions clearly explain the observed result.

关 键 词：函数数据分析 社会消费品零售总额 居民可支配收入 长期趋势 

分 类 号：F22[经济管理—国民经济]