Research On Vine Copula Modeling

The financial market is a fertile ground of the dependency research.In recent decades,the researchers have carried out a lot of research on the dependency of the financial series.Copula function,by virtue of its advantage in capturing tail relevance,is regarded by researchers as a powerful tool for studying multivariate dependent structures.However,Copula function's application has been limited by the computer technology of that era.However,with the development of computer technology and the rise of big data,the application and advance of Copula function are now available.But for the classic Copula function in the application there are still problems to be solved.Financial market is an enormous set of time series with complex structure.The use of classic high-dimensional Copula function will lead to a lack of complexity,along with the heavy burden of calculation and other issues.In this context,Vine Copula began to get its way in solving the problems of Copulas facing high-dimensional interdependent structures.Vine Copula decomposes the multivariate joint into a hierarchical structure by means of a graphical tool,'Vine',by which researchers can decompose high dimensional Copulas into a form of a product of several binary Copulas,reducing the computation needed.Such approach makes up the shortcomings of classic copula functions,which is of great theoretical significance.The Vine Copula is a theory of emerging interdependence in recent years.Although there is theoretical progress in Vine Copulas,many problems still need further discussion.In the structure identification process of the model,Vine Copula has more problems than the traditional Copula.As a basic work of choosing the appropriate Copula function,the determination of the Vine structure not only affects the fitting result of the whole Copula model,but also affects the researcher's point of inspective.The lack of comprehensive dependency indicators may cause the researcher to ignore some of the special dependent information between the sequences being studied.According to the matters mentioned above,this article focuses on the construction of the dependency measures used in the Vine Copula function's structure determination.On the basis of Vine Copula theory,this artical constructs the weighted rank correlation coefficient and the generalized correlation coefficient deduced in terms of the concept of rank correlation coefficients and generalized correlation coefficients,by which means an improvement of the sensitivity of the dependency measure in the Vine Copula modeling to the extreme events can be done,providing better reference measures for the process of the Vine Copula modeling.To show the performance of the new indicators in the modeling of the Copula model,an empirical test is conducted.The test result shows that the new indicator does enhance Vine Copula's ability of capturing tail dependence in extreme events,which have certain reference value for controlling financial risks.