非线性协整的非参数识别  

Nonparametric Identification of Nonlinear Co-integrated Systems

作　　者：马薇 葛通 肖凯 Ma Wei;Ge Tong;Xiao Kai(College of Economics & Management,Taiyuan University of Technology,Taiyuan 030024, China)

机构地区：[1]天津财经大学统计学院

出　　处：《统计与决策》2018年第24期10-14,共5页Statistics & Decision

摘　　要：非参数回归可以视作对非线性关联关系的机器穷举,把多种类型的非线性协整纳入同一框架内讨论,可使非线性协整研究更为简捷。为了避免过拟合所造成的协整检验取伪,文章提出了交错鉴定这一窗宽设定方法,使得识别更为稳健。详细讨论了非参技术在协整领域的应用细节,给出了多窗宽对比研究的数据分析思路。最后以汇率数据与出口数据为例,验证了方法,结果与理论相符。Nonparametric regression can be regarded as the machine exhaustion for nonlinear correlation, and it is easier to study nonlinear co-integration by bringing various kinds of nonlinear co-integration into one framework. In order to avoid co-integration forgery caused by over-fitting, this paper proposes a method of setting the window width of interlaced identification so as to make the identification more robust. The paper also discusses the application details of nonparametric technique in co-integration field, and offers the data analysis of multi-window width contrast research. Finally, the paper takes the exchange rate data and export data as an example to verify the method, with the results consistent with the theory.

关 键 词：非线性协整 非参数回归 过拟合 

分 类 号：F224[经济管理—国民经济] 0212