摘要
运用互相刺激的Hawkes过程研究中国股市暴涨和暴涨之间的交互作用。结果表明,在暴涨和暴跌幅度都服从广义帕累托分布的情形下,Hawkes过程能很好地拟合两者之间的相互作用。由模型可得,无论发生暴涨还是暴跌事件,都将显著地刺激下一个暴涨和暴跌的发生,这说明,中国股市体现出很明显的大波动聚集特征;此外,暴涨和暴跌都对同类事件的刺激持续更长时间。最后,运用该模型对中国股市未来发生暴涨和暴涨的时间进行相应预测。
We use mutually exciting Hawkes processes to study the interaction between large rises and large declines in China＇s stock market. Under the assumption that large rises and large declines follow the generalized Pareto distributions, the analysis shows that the proposed mutually exciting Hawkes process fits well the interaction between them. From the model, if either large rise or large decline occurs, it remarkably stimulate the occurrence of next large rise and large decline, which shows that large volatility clustering apparently exists in Chinese stock market; moreover, large rise and large decline have relatively a longer influence on the occurrence of the homogeneous events than that on the heterogeneous events. We also use the model to predict the occurrence times of next large rise and large decline.
出处
《系统管理学报》
CSSCI
2014年第5期755-760,共6页
Journal of Systems & Management
基金
国家杰出青年科学基金资助项目(70825005)
国家社科基金重大项目(11&ZD156)
关键词
大波动聚集
标记点过程
互相刺激
Hawkes过程
large volatility clustering
marked point processes
mutually exciting
Hawkes processes
