The Family Of SV Model Research And Application In China's Stock Market

Relevant to discuss and explore the volatility of the financial time series has been the core of the financial markets. About the volatility models research, researchers have line of sight gradually from the GARCH model to the SV model, and have proposed many of the extended models.Because it is difficult to estimate the SV model's parameters, so far most studies still remain in the basic SV model, but it's not satisfied for us. One side, compared with the normal distribution assumption, the financial time data's heteroscedasticity model will show a peak thick tail. Another reason is that there was a correlation between the rate of return and volatility of the data sequence. In order to solve the above problems, this thesis analyzes the stock market in our country by applying thick tail, mean value and leverage SV model on the basis of standard SV model.This thesis carries out Bayesian analysis for SV-T model, SV-MN model and leverage SV model in SV model by use of Bayesian theory, MCMC method and solves the posteriori distribution density in every model and constructs numerical calculation process based on Gibbs sampling. By use of Openbugs software, estimates in every model are obtained. After that, this thesis analyzes and then compares on behalf of the Chinese domestic stock market, the Shanghai Composite Index and on behalf of the Hong Kong stock market, the Hang Seng Index respetively through the obtained estimates. At last, this thesis carries out researches for the simulation situation of two indexes in every model by application of DIC principle in order to find out a model that fits the financial market in our country.