我国股票关联网络的动态演化研究  

Study of dynamic evolution of Chinese stock price correlation network

作　　者：黄玮强[1] 庄新田[1] 姚爽[2] 

机构地区：[1]东北大学工商管理学院,辽宁沈阳110819 [2]沈阳化工大学经济与管理学院,辽宁沈阳110142

出　　处：《系统工程学报》2014年第2期192-201,共10页Journal of Systems Engineering

基　　金：国家自然科学基金资助项目(71001022;71371044;71201108;71271047);中国博士后科学基金资助项目(20100471460);中国博士后科学基金特别资助项目(2013T60295)

摘　　要：利用最小生成树算法构建动态演化的我国股票关联网络.实证研究关联网络拓扑结构特征的动态演化规律、市场整体及个股价格行为与网络拓扑结构特征间的内在关系.结果表明:网络节点度和点强度服从无标度分布;随着时间的推进,股票间的价格波动关联关系呈现越来越强的稳定性;市场收益性越高及市场收益波动越小,网络平均距离就越大,股票间的价格波动关联性就越弱.较大的市场收益波动会导致网络变得更加收缩.绝大多数股票的收益率与其对应网络节点的度或点强度值呈负相关关系.市场中同行业的股票间易产生价格波动关联,验证了行业分类标准在投资组合构建中的参考价值.By using the minimal spanning tree algorithm, this paper constructs dynamic evolving Chinese stock price correlation networks. Then the dynamic evolution rules of network topology structures and the inner relationships between the market or individual stock＇s price behaviors and the network topology structures are empirically studied. The results demonstrate that the network degree obeys a scale free distribution. The price correlations get more and more stable with the network evolution. The higher the market return and the smaller the market return volatility, the larger the network＇s average path length, and the weaker the price correlation among stocks. A larger market return volatility will make the network shrink. Most of the stock returns have a negative relationship with the corresponding node degree or strength degree. The stocks within a same industry tend to have price correlations. It verifies the importance of industry classification standard in the portfolio construction.

关 键 词：股票关联网络 最小生成树 拓扑结构特征 收益性 波动性 

分 类 号：F830.91[经济管理—金融学]