基于随机矩阵理论的股市网络拓扑性质研究  

Study on Topological Property of Stock Market's Network Based on Random Matric Method

作　　者：谢赤[1,2] 胡珏 王钢金 

机构地区：[1]湖南大学工商管理学院,湖南长沙410082 [2]湖南大学金融与投资管理研究中心,湖南长沙410082

出　　处：《运筹与管理》2018年第1期144-152,共9页Operations Research and Management Science

基　　金：国家自然科学基金项目(71373072);国家自然科学基金项目(71340014);高等学校博士点专项科研基金项目(20130161110031)

摘　　要：本文运用随机矩阵理论(RMT)和相关系数动态演化模型建立全球股指二次"去噪"相关系数矩阵,并采用阀值法构建全球股市网络,进而分析该网络拓扑结构特性和解释风险在网络中的传染效应。研究发现,全球股市网络呈现出"小世界"效应;在θ=0.1数量水平下,全球股市网络具有较强的鲁棒性。同时,英国和荷兰的股票市场风险传染对网络整体的冲击较大;股市网络中各个股市间的风险传染路径与相关国家经济实力相关联,体现出较强的同配性。In this paper, we apply the random matric theory （RMT）and dynamic model of correlation coefficient to establish a worldwide stock markets＇ correlation coefficient matric of twice of denoising. We use the threshold method to build the stock markets＇ network, and further analyze the topological properties of the network and ex- plain the risk contagion among stock markets in the network. The study indicates that, the network is a small- world network. Under the threshold 19 = 0.1, the worldwide stock markets network is a strongly robust. As for risk contagion, the stock markets of England and Nederland have a strong impact on the stock market network than others. Besides, the risk contagion among stock markets in the network shows an assortativety.

关 键 词：股市网络 拓扑结构 风险传染 随机矩阵理论 

分 类 号：F830.9[经济管理—金融学]