Based On Symbolic Time Series Of Ultra-high Frequency Financial Volatility

The previous research results, the main use of financial data and other low-frequency high-frequency data interval, starting from the establishment of the ideological model to describe the characteristics of volatility, but modeling approach has its limitations, this article use Non-modeling approach, use ultra-high frequency data, describe the financial turmoil from a different angle.The first chapter describes the research background, I propose the thesis problem. The second chapter is literature review, including ultra-high frequency time series specific review of the studies, and Symbolic time series analysis literature review of specific studies, focusing on the study of ultra-high frequency time series modeling methods, and symbolic time series analysis Principle of the method.In applied research, firstly this article use three different symbols methods, analysis the changes of the trading interval of Industrial and Commercial Bank and Pudong development, the use of Shannon entropy of the improved Select the length of the string; of the three symbols of banking transactions were revealed differences in changes of the interval. Secondly, this article use the difference method, reveal the changes of number of transactions. Chapter IV put the interval and number of transactions on the same Cartesian coordinate system, describe the variation of the interval with subtraction method, and secondly, this article use the adaptive partition symbols, respectively divide the spacing to the trading interval and number of transactions. This article select representative stocks from different stock sectors, analysis the difference between the largest of the smallest trades and industries, and its Causes.