Selection And Comparison Of GARCH Family Models And SV Models

GARCH family models and SV models are very popular in financial time series analysis. This paper will comprehensively introduce and analyze these models. I will use the methods including the limiting conditional tail probability and the coefficient of tail dependence to distinguish and compare the GARCH models and the SV models. Then I introduce the approaches to select and compare GARCH models, GJR models and EGARCH models by testing method and the loss function method. Therefore, for different financial time series, one can always select the most suitable model through statistical analysis.Based on the conclusions of the tail characteristic of the GARCH models, I will study the limiting conditional tail probability and the coefficient of tail dependence of the GJR models. Finally, I will use the methods introduced in this paper to analyze the Shanghai composite index data and to choose the suitable model for the data. I also find the difference on leverage effect between Chinese emerging stock market and the mature financial markets.