Time-and state-domain methods are two common approa ches for nonparametric prediction.The former predominantly uses the data in the recent history while the latter mainly relies on historical information.The pa per adds the affect of noisy in time-domain,and obtains the volatility of nois y high frequency data by two time scales method,then get the dynamic integratio n volatility.%时间域和状态域方法是两种常见的非参数估计方法.前者主要使用的是最近的历史数据,而后者则主要依赖于过去的历史信息.本文在时间域上,通过对含噪音高频数据采用双时间尺度方法获得其波动率,进而获得经动态整合后的波动率.
展开▼