D-G正态模型在VaR计算上的改进  

The improvement of the D-G normal model on computing the VaR

作　　者：施建华[1] 黄可明[1] 

机构地区：[1]福州大学数学与计算机科学学院,福建福州350002

出　　处：《福州大学学报（自然科学版）》2004年第z1期20-23,共4页Journal of Fuzhou University(Natural Science Edition)

基　　金：福建省自然科学基金资助项目(E0310015)

摘　　要：对于包含期权等非线性头寸的投资组合来说,其VaR计算,通常是利用二阶或高阶泰勒展开式来近似投资组合在特定时期内相对于市场变量的价值变化,即D-G正态模型,然后针对这个模型来进行VaR计算.本文在分析这个模型缺陷的基础上,设法通过调整置信参数α,来提高风险预测的准确度.VaR methods based on a Delta Model are unlikely to be robust when applied to the portfolio including nonlinear positions. We commonly used a moment-matching method with the D-G Normal Model and calculated the VaR. In this paper, by analyzing the defect of the model, I tried to regulate the parameter, so as to raise the accurate level.

关 键 词：VAR 投资组合 D-G正态模型 期权 

分 类 号：O221.61[理学—运筹学与控制论]