Bayesian Analysis of Stochastic Volatility Models in Financial Time Series

摘要：

The stochastic volatility is a universal phenomenon in an economic and financial time series, and an important issue in the finance risk management research. In this paper,through the statistical structure of the standard stochastic volatility model, we infer the SV models likelihood function,design the parameters conjugate prior distribution, obtain the corresponding model parameter according to the Bayesian theorem, and examine their condition distribution. Furthermore, in order to obtain the model parameter estimation and their confidence intervals, we use Gibbs sampling to devise an MCMC computational procedure,and carried out an empirical analysis using the Shanghai composite index and the ShenZhen ingredient index data.The results indicate that the Bayesian method is an effective tool to explore the financial time series data.

关键词： Stochastic volatility models time series Bayesian method Gibbs sampling MCMC simulation Gibbs sampler

作者: Huiming Zhu Rui Zhao Liya Hao

作者单位: College of Business Administration Hunan University, Changsha 410082, China

会议类型: 国际会议

会议名称: 2007 Conference on Systems Science, Management Science and System Dynamics(第二届系统科学、管理科学与系统动力学国际会议)

会议地点: 上海

会议语种:英文

页码: 3195-3200

在线出版日期: 2007-10-19（万方平台首次上网日期，不代表论文的发表时间）