Construction Of Extreme-value Copulas Based On Spectral Measureand

In the real world, although the extreme events rarely occur, it will produce significant impact once happened. Extreme value theory (Extreme Value The-ory) researches modeling and statistical analysis method of extreme of random variables. Because of the complexity of the real world, extreme events more and more tend to occur at the same time or ensued, multivariate extreme value theory provides an effective method for solving this kind of problem.Constructing the multivariate extreme value distribution is an importan-t problem in multivariate extreme value theory. Copula plays an important role. Copula can be used for represent multivariate extreme value distribution. Construction of multivariate extreme value distribution is equivalent to the con-struction of extreme value Copula. According to maximum stability and spectral measure method of extreme value Copula, construct extreme value Copula by construction of tail dependence function. In this paper, introduce a method of constructing extreme value and two kinds of structure models:one is construct-ing extreme value with the same dimension copulas, in which the main research is the extreme value with discrete spectrum measurement. Two-dimensional dis-crete spectrum measurement is studied in detail, introducing the structure of the measure of the extreme value connect characteristic and its correlation coefficien-t. Second, we study the product structure of lower dimensional copulas connect high-dimensional extreme value connect. Not only we study the discrete spectrum measurement structure the high ones, and how to higher dimensional correlation function by low dimensional ones, and then construct of extreme value, researching its correlation to enrich the theory of extreme value Copula system.