Conditioning-based metrics on the space of multivariate copulas and their interrelation with uniform and levelwise convergence and Iterated Function Systems

被引:29
|
作者
Fernandez Sanchez, Juan [1 ]
Trutschnig, Wolfgang [2 ]
机构
[1] Univ Almeria, Grp Invest Anal Matemat, La Canada De San Urbano, Almeria, Spain
[2] Salzburg Univ, Dept Math, A-5020 Salzburg, Austria
关键词
Copula; Stochastic measure; Markov kernel; Iterated Function System; Level set; Endograph; SHUFFLES; APPROXIMATION; DEPENDENCE; SETS;
D O I
10.1007/s10959-014-0541-4
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
Using the one-to-one correspondence between copulas and special Markov kernels three strong metrics on the class C-rho of rho-dimensional copulas with rho >= 3 are studied. Being natural extensions of the two-dimensional versions introduced by Trutschnig (J Math Anal Appl 384:690-705, 2011), these metrics exhibit various good properties. In particular, it can be shown that the resulting metric spaces are separable and complete, which, as by-product, offers a simple separable and complete metrization of the so-called partial derivative-convergence studied by Mikusinski and Taylor (Ann Polon Math 96:75-95, 2009, Metrika 72:385-414, 2010). As an additional consequence of completeness, it is proved that the construction of singular copulas with fractal support via special Iterated Function Systems also converges with respect to any of the three introduced metrics. Moreover, the interrelation with the uniform metric is studied and convergence with respect to is characterized in terms of level-set and endograph convergence with respect to the Hausdorff metric.
引用
收藏
页码:1311 / 1336
页数:26
相关论文
共 2 条