摘要
对随机波动率模型的统计结构进行分析,基于贝叶斯定理,推导出模型参数的后验分布,利用MCMC算法对参数进行估计,同时将FFBS算法引入到波动率向量的估计过程中,对波动率序列进行联合抽样,提高Gibbs抽样算法的效率。对深圳基金指数和上证基金指数进行实证分析,结果表明:基于FFBS算法的随机波动率模型能很好地拟合数据的波动特征。
By analyzing the statistical structure of the stochastic volatility model, and using the Bayesian theorem,we derive the posterior distribution of model's parameter. We use the Markov chain Monte Carlo algorithm toestimate model's parameter, and the FFBS (forward filtering and backward sampling) algorithm to estimate thevolatility. We apply the stochastic volatility model to analyze the Shenzhen and Shanghai fund market, and theresult indicated that the model can explore the volatility characteristics of fund market.
出处
《佛山科学技术学院学报(自然科学版)》
CAS
2016年第3期5-10,共6页
Journal of Foshan University(Natural Science Edition)
基金
国家自然科学基金资助项目(11171117)