中文摘要：

利用格兰杰因果图表示多维时间序列变量间的因果关系,图中的点表示时间序列分量的点过程,图中的有向边反映序列间的格兰杰因果关系,无向边表示即期因果关系。利用部分定向相干技术研究了格兰杰因果图结构的识别问题,数值模拟表明该方法是有效的。将格兰杰因果图及识别方法应用于国际股票市场分析主要股指的信息流动,结果表明美国和香港股市在信息传导中起着重要作用,中国与国际主要股票市场的直接信息传导较弱,国际股市在信息流动中存在一定的区域效应。

英文摘要：

The causal relations among components of multivariate time are expressed by Granger causality graph. The vertices, representing the components of the time series, are connected by directed edges according to the Granger causality relations between the variables whereas undirected edges correspond to contemporaneous dependence. The structure identification of Granger causality graph is proposed based on partial directed coherence. The validity of the proposed method is confirmed by simulations. At last, the method is applied to the detection of information transmission in major international financial market. Empirical results show that there is the strong connection between US, HK and other stock markets, Chinese stock market is weekly connected among the major international markets. The Regional segmentation of the major international financial markets is proved in this study.