Stochastic Volatility Model Of Stock Market Return And Stochastic Simulation

This thesis is mainly for the skew-normal distribution of stochastic volatility model under scale mixtures under detailed analysis of the bias. Since the error distribution contains skewness and heavy-tailed distribution, such as Skew-t distribution, skew-slash distribution, the skew-contaminated normal distribution, this paper discusses a new model which provides an useful general symmetric nonlinear regression model. This type of distribution is a major disaster, is a performance is very good, which also allowes us to use Monte Carlo method to simulation the empirical distribution of samples with posterior.In order to stability test this flexible model, excluding the outliers and influential observations. In addition,we also studied some problems in model selection. And we use data from the Securities Composite Index and SZSE Composite Index under WINBUGS software in 2006-2014.