摘要
提出了新的基于体制转换模型的图模型方法.首先给出精度矩阵参数化方法,且通过MCMC方法给出算法设计,然后将该图模型方法应用于我国上海证券市场,研究五大行业板块之间的动态条件相关性.实证结果表明:两状态图结构恰好体现高、低不同的条件相关性,且状态持续的概率较大.
The Markov switch regime model for the structures of graphs is introduced and the algorithm by the Markov chain Monte Carlo method is proposed. The graphical model is applied to the stock market of Shanghai in China to study the conditional dynamic relativities of five segments of the market, Empirical results show that the two states reflect high and low of states is comparatively large. conditional correlation and the persistent probability of states is comparatively large.
出处
《系统工程理论与实践》
EI
CSSCI
CSCD
北大核心
2011年第5期881-888,共8页
Systems Engineering-Theory & Practice
基金
国家自然科学基金(10971042)
温州市科技计划项目(R20100030)
