基于可见图方法的股票行业分析

Stock Industry Analysis Based on Visibility Graph

复杂网络已被广泛应用于探究复杂系统的规律。本文使用可见图方法,分别将道琼斯工业指数30支成分股的日收盘价序列映射到复杂网络,对股票可见图的性质进行了分析,探究股票市场的网络结构的变化。结果表明,首先,股票原始序列可见图的度分布表现为幂律度分布,而随机打乱之后的序列可见图度分布呈指数分布;其次,可见图的网络属性可以反应出股票序列的波动情况。最后,网络属性的聚类分析可以识别行业领域相近的股票序列。可见图方法从宏观的角度揭示不同地区股票市场的性质和潜在动力学行为,能有效解析股票市场对外界信息的反映效率,反映了股票市场是以非线性的方式对外界信息做出反应。

Complex networks have been widely used to explore the laws of complex systems. This paper uses the visibility graph method to map the daily closing price series of 30 constituents of the Dow Jones Industrial Index to a complex network, to analyzes the characters of stocks’ visibility graphs, and explores the changes in the network structure of the stock markets. The results show that, firstly, the degree distribution of the visibility map of the original sequence of stocks is a power-law degree distribution, while the visibility degree distribution of the sequence after random disruption is ex-ponentially distributed. Secondly, the network properties of the visibility graph can reflect the fluctuations of the stock series. Finally, cluster analysis of network attributes can identify similar stock series in industry fields. The visibility graph method reveals the nature and potential dynam-ic behavior of the stock market in different regions from a macro perspective, which can effectively analyze the efficiency of the stock market's response to external information, and reflects that the stock market responds to external information in a non-linear way.