A Research Of Shanghai Composite Index Return Rate Based On ARFIMA Model

Our securities market is a weak efficient market. There are a large number ofstudies have shown that time series has the characteristics of long memory in weakefficient market. So we can’t ignore the long memory feature when analysis theShanghai Composite Index return rate. The ARFIMA model was select to analysisthe Shanghai Composite Index return rate which is from2008to2010in this paper.The long memory feature existence in the series has proved by R/S method. Bayesianmethod was used to analysis the ARFIMA model in this paper. The likelihood func-tion is established under t distribution. The prior distribution was sums up from otherestimation methods and the posterior distribution is simulated by MCMC method.The mean value of the parameters used as estimated value in this paper. By com-paring with the likelihood estimation. It shows that the bayesian estimation is betterthan the maximum likelihood estimation, but it still has large error. By analyzing thereasons of error, we put forward the prospects for further optimization at the end ofthis paper.