【摘要】 介绍了一种成熟的已广泛应用于金融领域的估计不确定性的方法——copula函数方法,并推广了n重的Frank’s copulas;在实际应用中,本文采用了俞言祥和美国西部由Boore等人给出的两个衰减模型,针对其中存在的模型不确定性以及它们之间的相互依赖性,构造出概率意义上联合的copula分布函数,并将其应用于实例分析.结果表明,对比于传统逻辑树中所用的线性结合方法,copula将两者带来的概率分布写成一个联合概率分布,能够很好地考虑双方不尽相同的意见.另外,由于copula函数可采用各种各样的边际分布函数来获得联合概率分布,且在金融风险投资评估已有大量的应用,因此在现代地震危险性评估中将有着广泛的前景.更多还原

【Abstract】 This paper introduces a new mature mathematical technique used to deal with models’ uncertainty in financial risk analysis-copula joint function, and generalize it to the n-dimensional Frank’s copula. In addition, we used a two-dimensional copula joint probabilistic function as an example to illustrate the uncertainty treatment at low probability level between two different attenuation models derived from Yu and Boore et al, and the result shows a better estimation than that given by linear weight technique usually used in the traditional logic-tree method. In light of widespread application in the risk analysis in the financial investment and insurance assessment, we believe that the copula-based method will have a potential application in the seismic hazard analysis.更多还原