摘要
本文利用非参数贝叶斯方法进行随机波动建模。通常的参数随机波动模型适用于证券市场中的综合指数数据,而对个股数据和小范围指数数据的拟合效果较差,主要原因是其收益率数据的变化规律更为复杂、具有更厚的尾部行为,而非参数贝叶斯方法的随机波动模型无需进行分布假设,具有很强的灵活性。本文利用SV-DPM模型对IBM的股票价格数据和上证50指数数据进行建模,研究发现非参数随机波动模型能拟合参数随机波动模型难以扑捉到的数据特征,实证表明有充分的依据支持非参数贝叶斯随机波动模型。论文的研究有助于捕捉金融资产的时变波动性质,能更好的揭示金融市场的运行规律,为期权定价和金融风险管理提供依据,对于防范与控制金融风险有着重要意义。
In this paper, we use Bayesian nonparametric stochastic volatility modeling methods. Parametric stochastic volatility model is applicable to the stock market composite index data, individual stock price data and small-scale index data fitting result is not very good. Instead of parametric Bayesian model without distributional assumptions, it can learn directly from the probability distribution of the data,it is very flexible. We used the SV-DPM model to fit IBM’s stock price data and the SSE 50 index.We found that non-parametric stochastic volatility model can found the data characteristics that the common parameter model is difficult to capture. Empirical evidence shows that there is a sufficient basis to support Bayesian nonparametric stochastic volatility model. The study enriched the discrete time stochastic volatility modeling method, and it is also helpful to capture the time-varying volatility of financial assets and reveal the rules of financial markets for assets within the framework of option price,provide decision-making basis for control financial risks and has important theoretical and practical significance for the financial system.
作者
蒋远营
张波
JIANG Yuan-ying;ZHANG Bo(College of Science,Guilin University of Technology,Guangxi Guilin 541004,China;Center for Applied Statistics,Renmin University of China,Beijing 100872,China;School of Statistics, Renmin University of China,Beijing 100872,China)
出处
《数理统计与管理》
CSSCI
北大核心
2019年第1期49-61,共13页
Journal of Applied Statistics and Management
基金
国家自然科学基金(71471173)
教育部人文社科基地重大项目(14JJD910002)
广西高校科研重点项目(KY2015ZD054)
广西自然科学基金联合资助培育项目(2018JJA180058)
