Modeling Dependence with C- and D-Vine Copulas: The R Package CDVine
Modeling Dependence with C- and D-Vine Copulas: The R Package CDVine
复制标题
DOI：
10.18637/jss.v052.i03
复制
发表时间：
2013-02
影响因子：
5.8
通讯作者：
E. Brechmann;U. Schepsmeier
中科院分区：
文献类型：
--
作者：
E. Brechmann;U. Schepsmeier
许多领域都需要灵活的多元分布。然而，流行的多元高斯分布具有很大的限制性，无法解释不对称和重尾等特征。因此，现在使用联结函数进行依赖建模来解释此类模式非常常见。然而，在更高维度中，连接函数的使用具有挑战性，其中标准多元连接函数的结构相当不灵活。 Vine copula 克服了这些限制，并且能够通过受益于丰富多样的二元 copula 作为构建块来对复杂的依赖模式进行建模。本文介绍了 R 包 CDVine，它提供了用于规范 vine (C-vine) 和 D-vine copula 统计推断的函数和工具。它包含用于双变量探索性数据分析和双变量联结选择以及藤蔓中配对联结家族选择的工具。模型可以顺序估计或通过联合最大似然估计估计。还包括采样算法和图形方法。
Flexible multivariate distributions are needed in many areas. The popular multivariate Gaussian distribution is however very restrictive and cannot account for features like asymmetry and heavy tails. Therefore dependence modeling using copulas is nowadays very common to account for such patterns. The use of copulas is however challenging in higher dimensions, where standard multivariate copulas suffer from rather inflexible structures. Vine copulas overcome such limitations and are able to model complex dependency patterns by benefiting from the rich variety of bivariate copulas as building blocks. This article presents the R package CDVine which provides functions and tools for statistical inference of canonical vine (C-vine) and D-vine copulas. It contains tools for bivariate exploratory data analysis and for bivariate copula selection as well as for selection of pair-copula families in a vine. Models can be estimated either sequentially or by joint maximum likelihood estimation. Sampling algorithms and graphical methods are also included.