We propose a subsampling method for robust estimation of regression models which is built on classical methods such as the least squares method. It makes use of
We develop higher order accurate estimators of integrated volatility in a stochastic volatility models by using kernel smoothing method and using different weig