(1) Realized bipower variation (BPV) enables us to effectively extract robust information about the diffusive volatility against jumps. As an application of BPV,we will introduce an empirical analysis of non-parametric jump-detection using the statistics proposed by Lee and Mykland (2008). (2) We will propose a statistical hypothesis test for removing noise in covariance matrix based on the nature in which maximum eigenvalues of covariance matrix asymptotically follow Tracy-Widom distribution. (3) Japanese stock markets have two types of breaks, overnight and lunch, during which no trading occurs, causing an inevitable increased variance in estimating daily volatility via a naive RV. In order to perform a more stabilized estimation, we will modify Hansen and Lunde's (2005) weighting technique. As an empirical study, we will estimate optimal weights by using a particular approach for Japanese stock data listed on the Tokyo Stock Exchange, and then compare the forecast performance of weighted and non-weighted RV through ARFIMA model. The empirical results indicate that the appropriate use of on optimally weighted RV can lead to remarkably smaller estimation variance compared with the naive RV, in many series. Therefore a more accurate forecasting of daily volatility data is obtained. Finally, we will perform a Monte Carlo simulation to support the empirical result.