作　　者： ; ; ;

机构地区： 华南理工大学理学院

出　　处： 《数学的实践与认识》 2007年第11期92-98,共7页

摘　　要： 通过ARMAGARCH模型模拟出可转债收益率的分布,然后使用遗传算法的数据挖掘技术对此分布进行了有效的优化,结合Bootstrap算法将优化结果应用于VaR测度,得出了GAVaR模型下的风险值.对中国和台湾可转债市场进行实证研究,发现GAVaR模型压力测试和回顾测试的结果都优于历史模拟法等常用VaR模型,控制风险的能力适合可转债市场的需求. Firstly the distribution of Convertible Bond yield is simulated by ARMA-GARCH model, then the new distribution is further optimized by genetic algorithm, and the further result is used in the field of VaR measure with the Bootstrap algorithm, finally we get the risk value which we called the GAVaR model. This paper also contains the empirical research of the China and Taiwan market, we find the results of Press test and Back test is better than that of the common VaR models, and the risk management ability of this new method is more suitable to the Convertible Bond market.

关 键 词： 可转债 风险测度 遗传算法

领　　域： [理学] [理学]