Nonparametric Method And Change Point Analysis Of Statistical Modeling In Econometrics

Econometrics is one of the branches of economics which is the unification ofeconomic theory, statistics and mathematics .But now, it is statistical modeling andsensitivity analysis that is one of the most important subjects of econometrics. Since the economic problems are mostly nonlinear, the economic data changes arenonstationary; the traditional econometric models are established under the assumptionof a long-range steady relationship between variables. To solve the problem, foreignresearchers try to use the methods of nonparametric, nonlinear and nonstationaryanalysis. And in China it also has been paid much attention to. The theory of kernelestimation discussed in this paper is one of these. The paper systematically introducesthe content on kernel estimation; especially, the methods of kernel estimation ofregression analysis are discussed deeply. Under some conditions, the consistency ofkernel estimation is summarized in theory, and the estimation becomes more stable, thetraditional parameter method has no such merit .Empirical analysis has been done tocompare the results by the concrete economic case. The other content of the research is change point analysis, which is also widely usedin all kinds of fields. Considering the work has been done, change point analysis has gotlots of results, and new researches arise frequently. This paper has done some deeplytheoretic research of the stability of change point analysis from another point of view,and got the results accordingly. Change point analysis is introduced into econometricmodeling .It faces up to practice better, exploring the test problem on change in modelsin econometrics, investigating its economic meaning. The according conclusions areapplied to study the external trade in our country, which has theoretical and appliedvalues.