经济数据分析:一种基于数据的函数性视角的分析方法

Economic Data Analysis:an Approach Based on Functional Viewpoint of Data

在经济数据分析中,经常会见到融合时间序列和横截面两者的数据。虽然20多年来,计量经济学大量研究了具有时间和横截面两维特征的面板数据,并提出了许多行之有效的分析方法,但是面板数据只是函数性数据的一种类型,且其分析方法太过于依赖模型的线性结构和假设条件等。本文基于经济数据的函数性特征,引入一种对数据进行分析的全新方法,并利用基于MATLAB编写的程序,率先对经济函数性数据进行分析,拓展了函数性数据分析的应用范围,填补了我国在此方面研究的空白。实例分析结果表明,函数性数据分析方法,较之计量经济学和其它统计方法具有更多的优越性,尤其能够揭示其它方法所不能揭示的数据特征。

Data set consisting of time series and crosssectional data can often be found in economic data analysis.Though numerous effective methods to study the panel data have been developed in econometrics during the last two decades,the panel data consist of a type of functional data and the methods rely heavily on the linear structure and assumptions.This paper presents a new data analysis method based on the functional features of data and we firstly use it to analyze economic data by exploiting the computer program based on MATLAB.This paper broadens the application scope of functional data analysis and fills the gaps in this aspect in China.The results show that the functional data analysis are more advantageous than econometrics and other statistical methods due to the fact that the former can reveal data features while the other methods can not.