正态分布和t分布下风险价值计算的对比研究  

Compared calculation of Value-at-Risk under normal and student distribution

作　　者：谷伟[1] 万建平[1] 王丽丽[1] 

机构地区：[1]华中科技大学数学系,湖北武汉430074

出　　处：《华中科技大学学报（自然科学版）》2004年第10期108-110,共3页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目 (10 1710 35 )

摘　　要：针对金融数据的尖峰、厚尾性 ,先在单变量情形下用经验贝叶斯方法得出正态分布和t分布下风险价值 (VaR)的计算公式 ,然后探讨了当金融数据服从多变量分布时 ,应用delta gamma二次近似法 ,分别得出正态分布和t分布下的组合VaR计算方法 。Aiming at financial data with high kurtosis and heavy tail, first we applies Empirical-Bayesian(EB) techniques to the normal and student models in order to obtain Value-at-Risk(VaR) under univariate case. The calculation methods such as delta-gamma quadratic approximation method of portifolio VaR under normal and student models were separately explored when the financial data were the multivariate distributions. The basis for solving heavy-tailed problem was provided.

关 键 词：风险价值 正态分布 T分布 delta-gamma二次近似法 

分 类 号：O212[理学—概率论与数理统计] F830[理学—数学]