多元copula参数模型的选择  

Choice of the Multivariate Copulas with Parametric Structure

作　　者：吴娟[1] 刘次华[1] 邱小霞[1] 文婷[1] 万芳[1] 

机构地区：[1]华中科技大学数学与统计学院,湖北武汉430074

出　　处：《武汉大学学报（理学版）》2008年第3期267-270,共4页Journal of Wuhan University:Natural Science Edition

基　　金：国家自然科学基金资助项目(10301011)

摘　　要：多元copula参数模型能完整刻画变量间的相关性关系.讨论了一类copula模型的选择问题,其多元copula函数能与一个一元函数构成一一对应的关系.对于Gumbel,Clayton,Frank,Fréchet等4种此类copula模型,分别在参数已知或未知两种情况下进行了拟合优度检验,并对中国股市的上证指数与深证综指作了实证分析,结果表明两者存在着较强的正相关性,相关性模型选取Gumbel copula模型最合适.The multivariate copulas with parametric structure can describe fully the dependence between variants. A special kind of multivariate copulas which can be only decided by a unary function is discussed. Goodness-of-fit tests are given to such four copula models as Gumbel, Clayton, Frank and Frechet, under both circumstances when parameter is known or unknown. We make empirical analysis of China＇s stock market. The results show that the Shanghai index and Shenzhen composite index share a strong positive correlation, and Gumbel is the most suitable copula in the dependence models.

关 键 词：COPULA函数 拟合优度检验 相关性关系 

分 类 号：O211.3[理学—概率论与数理统计]