论文部分内容阅读
传统的copula模型在对二维以上相依结构建模时存在参数过少的缺陷,vine copula理论基本弥补了这一缺陷.介绍了vine copula理论以及其相对于传统多元模型的优势,尤其提出了vine copula对于时长不一致的数据进行建模具有数据利用率较高的特性,给出了这类数据vine copula的建模步骤以及基于极大似然估计的统计推断.最后对国内A股市场的五种金融股票的联合分布进行建模,并利用蒙特卡罗方法对资产组合的VaR进行了模拟.
The traditional copula model has too few parameters when modeling the dependent structure of two dimensions or more, and the theory of vine copula makes up for this defect.The theory of vine copula and its advantages over the traditional multivariate model are introduced, especially vine copula modeling data with inconsistent duration has the characteristics of high data utilization, and gives the modeling steps of this kind of data vine copula and the statistical inference based on the maximum likelihood estimation.Finally, five kinds of domestic A-share market The joint distribution of financial stocks is modeled and the VaR of the portfolio is simulated using the Monte Carlo method.