采用MCMC方法的上海股市随机波动模型

采用贝叶斯统计中的马尔科夫链-蒙特卡罗(MCMC)方法对上海股市的随机波动性进行研究,基于Gibbs抽样的MCMC数值计算过程,对上海股市的随机波动率模型(SV)进行参数估计,并在WinBUGS软件中实现.根据信息判别准则(DIC),对比拟合的SV-N,SV-T,SV-MT模型参数,结果表明:SV-T模型最能反映上海股市波动具有尖峰厚尾的特性,可进一步用于预测样本外的波动率结果.

Stochastic Volatility Modeling of Shanghai Stock Exchange Using MCMC Method

One method is by Markov chain Monte Carlo(MCMC)bias statistics method. In this paper, we study the stochastic volatility of Shanghai Stock market, and estimate the parameters of the stochastic volatility model(SV)of Shanghai Stock market based on the MCMC sampling, and implement the Gibbs software in the WinBUGS software. By comparingthe parameters of SV-N, SV-T, SV-MT model, and according to discriminative information criterion, we find the SV-T model is the best model in China reflecting the fluctuation of the stock market of Shanghai which has peak thick tail characteristics, this model can also be used to step out of sample forecasting volatility results.