摘要
本文提出了两种风险值的估计方法,这两种方法均是先估计出收益的分布,然后求得分布左侧p分位点作为风险值的估计．第一种方法是用核估计方法得到收益的分布估计；第二种方法则是由分布的核估计算得收益的众数,引入所谓的广义半t分布拟合众数左侧的样本．文章以上证指数为实例验证了这两种方法的可行性与精确性．最后我们利用上述两种估计方法得到了上证指数风险值的波动主周期．
In this paper, two kinds of VaR estimate methods are put forward. Both of the methods are to estimate the distribution of the rate of return firstly, and then to calculate the estimated distribution＇s lower p critical value as the VaR. The first method for estimating distribution is kernel estimate method; In the second one, a general semi-t-distribution is defined to fit the rate of return data locating the left side of its mode. This mode is obtained from the distribution estimated by the first method. We take the SSE Composite Index as an example to demonstrate the feasibility and the accuracy of these two estimate methods. At last we find out the principal period of the VaR fluctuation of SSE Composite Index by above two estimate methods.
出处
《应用数学与计算数学学报》
2006年第1期82-86,共5页
Communication on Applied Mathematics and Computation
关键词
风险值(VaR)
核估计
厚尾
周期
Value-at-Risk （VaR）, kernel estimate, heavy tail, period
