分离估计归因于跳部分和连续部分对资产定价是非常重要的.由于市场信息的流入,前者通常比后者缺少可预测性.到目前为此,小波方法对于发现跳点和估计跳大小是有力的,正如王亚珍[1].但是在一些程度上,在点估计方面不能准确地对跳的位置和大小进行估计.本文中,我们提出了改进方法去估计已实现方差,从而新的估计量被用于在不同的取样策略和检验标准下估计跳平方和,进一步我们做了大量的模拟对有限样本性质进行验证.
Separating estimation due to jump and continuous parts are very impotent for asset pricing,and the former is usually less predictable than the latter due to the inflow of market news.So far,wavelets methods are powerful for detecting jump locations and jumps as demonstrated in Wang,Y[1].But to some extent,the jumps are not precisely estimated for locating and detecting in the pointwise predication.In this paper,we propose the adaptive method for estimating the realized variance,thus the new estimators are used to estimate the squared sum of jumps for the various sample strategies and test norms.Furthermore,we perform extensive simulations to examine the finite sample properties.