SV族模型的研究及在中国股市中的应用
发布时间：2018-09-13 08:01
【摘要】：有关对金融时间序列波动性的讨论及探究一直是金融市场研究的核心问题之一。对于波动模型的研究,近年来研究者们的视线逐渐从GARCH类模型转向各种扩展SV模型,根据研究金融市场的角度不同研究者们相继提出了众多的扩展模型。 由于SV族模型中参数很难估计,所以目前为止研究最多是基本SV模型,但是把它用于研究我国的股市,出来的结果不是很令人满意,因为相对于正态分布的假设,金融时间序列数据的异方差模型会呈现高峰厚尾性；另一原因是数据序列的收益率和波动性之间还存在着相关性,为了解决以上问题,本文在标准SV模型的基础上着重运用厚尾、均值以及杠杆SV模型对我国的股市进行了分析。 运用贝叶斯原理及MCMC方法对sV族模型中的SV-T模型、SV-MN模型以及杠杆SV模型进行了贝叶斯分析,求出每个模型中参数的后验分布密度,构造基于Gibbs抽样的MCMC数值计算过程,然后利用Openbugs软件求出模型中各参数的估计值,之后通过求出的估计值对代表中国国内股市的上证综指和代表香港股市的恒生指数进行分析并做比较研究,最后运用DIC准则对两股指在各模型下的模拟情况进行研究分析,找到较为适合我国金融市场的模型。
[Abstract]:The discussion and exploration of the volatility of financial time series has been one of the core issues in the study of financial markets. For the research of volatility model, researchers have gradually shifted their attention from GARCH model to various extended SV model in recent years. According to the perspective of financial market, many extended models have been put forward by different researchers. Because the parameters of the SV family model are difficult to estimate, so far the basic SV model is the most studied, but the result of applying it to the stock market in our country is not very satisfactory, because of the hypothesis of normal distribution. The heteroscedasticity model of financial time series data shows a peak and thick tail; another reason is that there is a correlation between the return rate and volatility of the data series, in order to solve the above problem, Based on the standard SV model, this paper analyzes the stock market in China by using the model of thick tail, mean value and leveraged SV. Bayesian analysis of SV-T model SV-MN model and lever SV model of sV family model is carried out by using Bayesian principle and MCMC method. The posterior distribution density of parameters in each model is obtained, and the MCMC numerical calculation process based on Gibbs sampling is constructed. Then the Openbugs software is used to calculate the estimated values of each parameter in the model, and then the Shanghai Composite Index, which represents the domestic stock market of China, and the Hang Seng Index, which represents the Hong Kong stock market, are analyzed and compared through the estimated values. Finally, the DIC criterion is used to study and analyze the simulation of the two stock indexes under each model, and to find a more suitable model for China's financial market.
【学位授予单位】：山东理工大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F832.51
本文编号：2240548
[Abstract]:The discussion and exploration of the volatility of financial time series has been one of the core issues in the study of financial markets. For the research of volatility model, researchers have gradually shifted their attention from GARCH model to various extended SV model in recent years. According to the perspective of financial market, many extended models have been put forward by different researchers. Because the parameters of the SV family model are difficult to estimate, so far the basic SV model is the most studied, but the result of applying it to the stock market in our country is not very satisfactory, because of the hypothesis of normal distribution. The heteroscedasticity model of financial time series data shows a peak and thick tail; another reason is that there is a correlation between the return rate and volatility of the data series, in order to solve the above problem, Based on the standard SV model, this paper analyzes the stock market in China by using the model of thick tail, mean value and leveraged SV. Bayesian analysis of SV-T model SV-MN model and lever SV model of sV family model is carried out by using Bayesian principle and MCMC method. The posterior distribution density of parameters in each model is obtained, and the MCMC numerical calculation process based on Gibbs sampling is constructed. Then the Openbugs software is used to calculate the estimated values of each parameter in the model, and then the Shanghai Composite Index, which represents the domestic stock market of China, and the Hang Seng Index, which represents the Hong Kong stock market, are analyzed and compared through the estimated values. Finally, the DIC criterion is used to study and analyze the simulation of the two stock indexes under each model, and to find a more suitable model for China's financial market.
【学位授予单位】：山东理工大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F832.51
【参考文献】
相关期刊论文 前5条
1 白],张世英;扩展SV模型及其在深圳股票市场的应用[J];系统工程;2001年06期
2 孟利锋,张世英,何信;具有杠杆效应SV模型的贝叶斯分析及其应用[J];系统工程;2004年03期
3 王春峰,蒋祥林,李刚;基于随机波动性模型的中国股市波动性估计[J];管理科学学报;2003年04期
4 黄大海,郑丕谔;股市预期收益率与波动关系的研究[J];管理科学学报;2005年05期
5 孟利锋,张世英,何信;厚尾SV模型的贝叶斯分析及其应用研究[J];西北农林科技大学学报(社会科学版);2003年06期
相关硕士学位论文 前3条
1 钱春沁;MCMC理论及其在金融资产定价中的应用[D];南京理工大学;2003年
2 李峰;基于MCMC模拟的贝叶斯金融随机波动模型分析[D];湖南大学;2007年
3 刘志丹;基于贝叶斯分析的厚尾和杠杆SV模型对中国股市的研究[D];南京理工大学;2009年
，本文编号：2240548
本文链接：https://www.wllwen.com/guanlilunwen/zhqtouz/2240548.html
最近更新
教材专著
