摘要
准确测度风险值VaR对投资组合选择及金融风险控制等提供了重要参考标准。Copula函数广泛用于VaR的计算,但在边缘分布建模参数、非参数及半参数方法的使用中存在较强的主观性。为此,提出了一种混合参数和非参数的金融资产边缘分布的半参数Copula建模方法,能将边缘分布的参数、非参数及半参数方法有机结合起来,并利用分布函数误差平方和最小准则来选择最优的资产分布模型。通过实证分析将其应用于资产投资组合的VaR计算中,并通过稳健性检验等方法进一步验证了该方法的有效性。
Accurately measuring VaR provides an important reference standard on portfolio selection and financial risk con- trol. While Copula have been widely used in the calculation of VaR, however, there are still strong subjectivity in the marginal distribution modeling. This paper presents a semi - parametric Copula modeling method involving parametric and non - parametric financial assets distribution, which can combine the parameters, non -parametric and semi -parametric methods above, and the minimum distribution function square error criterion is used to choose the optimal distribution model. Finally, after an empirical study of asset portfolio VaR calculation, robustness tests verify the validity of the new method.
出处
《武汉理工大学学报(信息与管理工程版)》
CAS
2016年第2期192-196,共5页
Journal of Wuhan University of Technology:Information & Management Engineering
基金
国家自然科学基金项目(71201131)
中国博士后科学基金项目(2014M562334)
成都市软科学研究资金项目(2014-RK00-00024-ZF)