›› 2010, Vol. 49 ›› Issue (增刊01): 65-69.

• 论文 • 上一篇    下一篇



  1. (1.山西忻州师范学院数学系,山西 忻州 034000;2.山西财经大学管理科学与工程学院,山西 太原 030006)
  • 收稿日期:2010-03-16 修回日期:1900-01-01

Use Complex Network to Research the Strong Correlation of Stocks in Some Sectors of Market

LAN Wangsen1,ZHAO Guohao2


  1. (1. Department of Mathematics, Xinzhou Teachers University, Xinzhou 034000, China;2. School of Management Science and Engineering, Shanxi Finance and Economics University, Taiyuan 030006, China)
  • Received:2010-03-16 Revised:1900-01-01

摘要: 为探索股票之间相互影响的行为,提高投资组合构建能力,以中国股市煤炭、电力板块股票为节点,以近19年股票对数回报的相关系数为边,建立复杂网络模型。通过对网络拓扑参数计算,发现该网络为无尺度网络,节点度分布负幂指数小于1,无权网络和加权网络平均集聚系数分别为0.68和0.41。对网络中心性进行了测量,发现000723,601898,601918三个节点是整个网络的核心节点;网络可划分成两个分区,并抽取出一个高度耦合的具有13个节点的中心网络,对整体网络有很大影响。

关键词: 复杂网络, 股票市场, 拓扑, 相关系数

Abstract: To explore the interaction among stocks and improve the ability to build a portfolio, a complex network is modeled in Coal and Power Sectors in Chinas Stock Markets, in which nodes are stocks, edges are correlation coefficients of stocks logarithm returns nearly 19 years. By estimating network topology parameters, some characters are found that networks scalefree, negative exponents of nodedegree distributions less than 1, average cluster coefficients 0.68 for unweighted networks and 0.41 for weighted networks. Networks centrality is measured, find that nodes 000723, 601898, 601918 are the core of the entire network. The network can be divided into two partitions, contracted into 2 center networks, one of them with 13 strongcoupling nodes to influent entire network greatly.

Key words: complex networks, stock market, topology, correlation coefficient