价格统计中大数据应用基础和总体思路

Application Bases and General Ideas of Big Data in Price Statistics

随着扫描数据和网上交易数据日渐普遍,在价格统计中应用大数据的热度也日渐高涨。以扫描数据的应用为基础,本文首先梳理价格统计中应用大数据的一些基础工作,其次讨论基础价格有关问题,然后系统思考价格指数理论,指出在动态数据集基础上直接通过双边指数短期链乘生成长期指数可能由于大数据高频、高搅动而存在链漂移,从而无法兼顾准确测度长期价格变动,进而明确在大数据环境下编制多边指数或使用动态方法生成指数的适用性和合理性。最后,考虑到大数据集的特点、划分同质可比产品集的优点、单位价值法的应用等前提,结合统计实践以及对指数理论和方法整体上的把握,分别从基础工作层面、基础价格层面、指数编制层面、编制CPI等综合指数的总体思路层面得出大数据应用的相关启示。

As scanner data and online transaction data are becoming more common,the popularity of big data in price statistics has also increased.Based on the application of scanner data,this paper sorts out basic works of using big data in the price statistics,and then discusses related issues of basic price,and then systematically considers the price index theory,pointing out that the long-term index generated through the short-term bilateral index chain on the basis of dynamic data set may have chain drift due to the high frequency and high churn of big data,so that the long-term price changes aren't available for accurate measurement,thus clarifies the applicability and rationality of the multilateral index or the use of a dynamic method to generate a index in the big data environment.Finally,considering the characteristics of big data set,the advantages of dividing homogeneous items,the application presumption of unit value method,and combined with statistical practice and the overall grasp of index theories and methods,we can get inspirations of big data application for the basic works side,basic price side,index compilation side,the comprehensive index(e.g.CPI)compilation side respectively.