极值指标下平稳序列的风险测度  

Risk Measurement of Stationary Sequence Under Extreme Value Index

作　　者：蔡宗鹏[1] 陈守全[1] CAI Zong-peng;CHEN Shou-quan(School of Mathematics and Statistics,Southwest Universit)

机构地区：[1]西南大学数学与统计学院,重庆400715

出　　处：《西南大学学报（自然科学版）》2018年第7期79-84,共6页Journal of Southwest University(Natural Science Edition)

基　　金：国家自然科学基金项目(11571283)

摘　　要：平稳金融序列存在相依性,不满足独立同分布假设下极值理论的条件约束.讨论了极值指标下平稳随机序列的风险测度,通过估计极值指标θ并修正广义极值分布的位置参数和尺度参数,得到平稳随机序列的风险值.实证和检验表明平稳金融序列需要极值指标修正模型,提高风险值估计的准确度.The stationary financial sequence is not independent while it does not satisfy the requirement of the conditional constraint of the extreme value theory under the consumption of modified independent identical distribution（i.i.d.）.In this paper,we discuss the risk measurement of the stationary random sequence under the extremal index,and derive the valve at risk of the stationary random sequence by estimating the extremal index and modifying the position parameter and the scale parameter of the GEV.Empirical and test results show that the stationary financial sequence needs the extremal index to improve the accuracy of the VaR estimation.

关 键 词：极值理论 平稳序列 相依性 极值指标 

分 类 号：O212.3[理学—概率论与数理统计]