基于符号序列分析的股市网络结构及金融波动分析-analysis of stock market network structure and financial fluctuation based on symbolic sequence analysis.docx

ABSTRACT With the development of national economy, the relationship between Chinas economic and the rest of the world’s ties more and more closely, at the same time, our economy is more and more influenced by the global financial market volatility. It is of great importance for realizing and spreading financial risk to analyze and forecast the volatility of financial market with scientific method. Abundant achievements have been made in the research of financial fluctuation with financial metrology method. However the economic system is a complicated nonlinear system, the rule of the financial volatility cant be grasped in all aspects just with the financial metrology method. Thus, new research perspective and method need to be put forward. As a supplement of the financial metrology research method, symbolic time series analysis method is tried to be introduced in the analysis of financial network and forecasting. Symbolic time series analysis method is introduced in this thesis, distance matrix is gotten based on the symbolic series’ coding series. On the basis of the above, the minimum spanning tree and hierarchical tree are gotten. And the analysis of stock market network structure is done according to that. Empirical analysis based on the stocks of Shanghai and Shenzhen 300 index is carried out. In the aspect of financial volatility forecasting, new method is introduced based on symbolic time series analysis and sequence alignment method. The best matching pattern is found by symbolizing the time series，adopting the appropriate pattern length and sequence allocating. The predictive value can be gotten based on the matching pattern. And empirical analysis is made with high frequency data whose sampling interval is 20 minutes from Shanghai Stock Exchange Composite Index. The first chapter of this thesis states the background, significance and research status of the relative research methods, combs the structure of this thesis and puts forward innovative points. T