基于多维灰色组合模型的CPI预测应用研究  

Application Research on CPI Prediction Based on Multi-dimensional Grey Combination Model

作　　者：王旻燕 王兴龙[1] 汪亚 WANG Minyan;WANG Xinglong;WANG Ya(Anhui Xinhua University,Hefei Anhui 230088)

机构地区：[1]安徽新华学院,安徽合肥230088

出　　处：《软件》2025年第2期37-39,共3页Software

基　　金：安徽省高等学校省级质量工程项目(2022jyxm673);安徽新华学院校级科研项目(2023zr025)。

摘　　要：CPI能够反映一定时期内人们购买一组代表性商品和服务总花费的变化情况。为提高多维灰色模型对区间数序列预测的精确度,本文将两个多维区间数序列灰色模型MINGM(0,2)、MINGM(1,2)进行组合。首先,分别建立两个单一模型并根据二者自身的拟合及预测结果确定变权系数,从而建立新的多维灰色组合模型;其次,用我国2008—2024年CPI数据作为系统特征序列,将衣着类居民消费价格指数作为相关因素序列,对组合模型的准确性进行评价。结果表明,改进后的多维灰色组合模型不论是拟合还是预测精度都得到了显著提升。CPI can reflect the changes in the total spending of a group of representative goods and services over a certain period of time.To improve the accuracy of multi-dimensional grey model in predicting interval data sequences,this paper combines two multi-dimensional interval grey models,MINGM(0,2)and MINGM(1,2).Firstly,two single models are established respectively and the variable weight coefficients are determined according to their own fitting and prediction results,so as to establish a new multi-dimensional grey combined model;and then,the accuracy of the combined model is evaluated by using the CPI data of China from 2008 to 2024 as the system feature series and the Consumer price index of clothing as the correlation factor series.The results show that both the fitting and prediction accuracy of the improved multi-dimensional grey combined model have been significantly improved.

关 键 词：多维灰色模型 区间数序列 变权组合预测 CPI 

分 类 号：O212.1[理学—概率论与数理统计]