摘要
经典的极值模型要求数据是独立同分布的,本文考虑平稳序列,引入极值指标,并利用分串方法,建立POT模型,对VaR和CVaR进行估计,最后对日元/美元的汇率进行实证研究。通过比较发现,引入极值指标后,提高了VaR估计的精度。
The classic extreme value theory requests that sequence is independent and has identical distribution. We intro- (duce) the extremal index under the assumption that the sequence is stationary, and build a POT model by using the method of declustering, then calculate the estimates of VaR and CVaR. The computation result of JPY/USD foreign exchange rate presented at last proves that the accurate for the estimations has been improved by introducing the extremal index.
出处
《系统工程》
CSCD
北大核心
2004年第6期49-53,共5页
Systems Engineering
基金
南开大学-天津大学刘徽应用数学中心资助项目(T08)
关键词
POT模型
极值指标
风险价值
条件风险价值
The POT Model
The Extremal Index
Value-at-Risk
Conditional Value-at-Risk
