Research On Financial Volatility Based On Empirical Mode Decomposition And Symbolic Time Series Analysis

The empirical mode decomposition method is introduced in the research of financial volatility,combined with symbolic time series analysis method,based on different characteristic time scales on the Shanghai Stock Exchange and Shenzhen stock exchange fluctuations in the financial situation to do comparative analysis,and then combined with BP neural network prediction model on the Shanghai and Shenzhen two city realized volatility trend analysis and forecasting.This thesis firstly describes the financial fluctuations from the background,significance and results of the three aspects,and then put forward the two research methods,symbol time series and empirical mode decomposition,summarized the research results of the two.Then under the background of symbolic time series,combined with empirical mode decomposition method,the Shenzhen Composite Index and Shanghai Composite Index of the realized volatility series of empirical mode decomposition and reconstruction,extract the short-term fluctuations in the two markets,long-term fluctuation and fluctuation trend,provide the basis for different types of financial market investors.Also in the different time scales based on symbolic sequence statistical method on the Shenzhen Composite Index and the SSE Composite index the difference analysis,thus to investors and regulators provide more considerations.At the same time,the BP neural network is introduced,combined with the EMD decomposition,and the regularity of the historical symbols is decomposed,to predict the future trend and risk level,and provide a new perspective and method for the prediction and research of financial market.Through the empirical analysis,the multi-scale phenomenon in the financial markets is found,and the difference of the two cities in the different time scales is found.The empirical mode decomposition method and BP neural network are combined with the prediction method of BP neural network.It is proved that the empirical mode decomposition has certain advantages in dealing with nonlinear and non stationary time series.