摘要
文章从VaR方法的定义出发,首先对VaR值的两种基本计算方法进行阐述,进而基于核密度估计,提出一种改进的VaR值计算方法。该改进方法将蒙特卡罗模拟法引入到核密度估计规则,并且考虑四分位距来构造核密度估计的窗宽,对股市收益率的变异性以及高峰厚尾现象进行了更好地刻画。实证验证了改进的VaR值计算方法的有效性及优越性。
Starting from the definition of VaR method, this paper firstly elaborates the two basic VaR calculation methods, and then proposes an improved VaR method based on kernel density estimation. This revised method introduces the Monte Carlo simulation method into the kernel density estimation rule, and constructs the window width of the kernel density estimation by us- ing quartile distance, which better characterizes the variability of the stock market returns and the phenomenon of leptokurtosis and fat-tail. Empirical verification shows the validity and superiority of the improved VaR value calculation method.
出处
《统计与决策》
CSSCI
北大核心
2017年第16期32-35,共4页
Statistics & Decision
基金
国家自然科学基金资助项目(11401591)
中央高校基本科研业务费专项资金资助项目(2014143)
关键词
VaR值计算方法
核密度估计
蒙特卡罗模拟法
四分位距
VaR value calculation method
kernel density estimation
Monte Carlo simulation method
quartile distance
