Analysis And Empirical Research Of High-frequency Of Financial Data

With the development of computer technology and innovation of communication technology, people could get and deal with financial high-frequency data easily. Since high-frequency data contained plenty of market information, extracting effective information from the data timely became more and more important.The paper analyzed high-frequency data of Chinese stock markets mainly from periodicity and long memory these two aspects. First, we did preliminary statistical analysis on high-frequency data of Chinese stock markets by traditional method, the results showed that the high-frequency return series of Chinese stock markets were not normal distribution, but peak and heavy tail, and had significant "calendar effect". Second, the paper introduced realized volatility and its expanded forms, and weighted realized volatility is more efficient than adjusted realized volatility and realized volatility by empirical analysis, it was because that weighted realized volatility fully considered "calendar effect", then, we had that the optimal sampling frequency is5min of weighted realized volatility of Shanghai composite index in2011. Last, it was proved that weighted realized volatility had long memory property both by R/S analysis method and modified R/S analysis method respectively, then, based on HAR-RV model, the paper proposed HAR-WRV model, and did empirical analysis on5min of weighted realized volatility and realized volatility of Shanghai composite index in2011, as a result, the two models both could fit and forecast volatility efficiently, and the fitting and predictable effect of HAR-WRV model was better than HAR-RV model.