基于网络搜索数据监测的CPI预测  

CPI prediction based on data monitoring of Internet search

作　　者：闫妍[1] 张志欣 张洪瑶 YAN Yan;ZHANG Zhixin;ZHANG Hongyao(School of Management,Shenyang University of Technology,Shenyang 110870,China)

机构地区：[1]沈阳工业大学管理学院,沈阳110870

出　　处：《沈阳工业大学学报（社会科学版）》2023年第4期365-370,共6页Journal of Shenyang University of Technology(Social Sciences)

基　　金：辽宁省教育厅面上项目(LJKR0078)。

摘　　要：网络搜索经济类关键词的范围和频率能反映公众的关注度和倾向,可用来对经济指标进行预测,从而预示经济走势。为解决居民消费价格指数(CPI)预测中的官方数据滞后问题,使用时间序列法,采用网络搜索监测数据对CPI进行预测。运用百度指数分析方法,搜集宏观、微观、投资三类指标,对初始指标的百度指数进行主成分分析,拟合出CPI预测模型。结果表明,预测CPI和实际CPI走势基本一致,可比官方数据提前一个月得到预测值,为把握宏观经济走势提供参考。The scope and frequency of Internet search for economic keywords can reflect the attention and tendency of the public,and can be used to predict economic indicators,so as to predict the economic trend.In order to solve the lag problem of the official data in the prediction of Consumer Price Index(CPI),the time series method is used to predict CPI by applying monitoring data of Internet search.Applying Baidu Index analysis method,three types of indicators are collected of macro-,micro-and investment-indicators,principal component analysis is conducted on the Baidu Index of initial indicators,and the CPI prediction model is fitted.The results show that the predicted CPI is basically consistent with the actual CPI trend,and the predicted value can be obtained one month ahead of the official data,which provides reference for grasping the macro-economic trend.

关 键 词：宏观经济 经济走势 网络搜索 居民消费价格指数 数据监测 百度指数 

分 类 号：F201[经济管理—国民经济]