Research on Stock Network Modeling and Risk Propagation Based on Complex Network Analysis

The price fluctuation in the stock market is often accompanied by risks, the spread of which affects the entire financial system and even the stable operation of the social economy as a whole. Considering the relatively mature research for the network characteristics of the stock market, this paper attempts to employ the complex networks analysis method to study the risk of the stock market. We firstly analyze the determinants of the stock influence (SI), put forward the measurement method of SI and thus establish the risk propagation model. On that basis, we propose GASI, the influence maximization algorithm for stock network, and dig out the key nodes that maximize SI. Experimental results show that price fluctuations of a small number of stocks may largely impact the other stocks, even the entire stock market. These influential stocks distribute in almost all industries, but are mainly in finance, manufacturing and mining. In addition, compared to the PageRank algorithm, the GASI algorithm can influence more nodes and widen the range of risk propagation, and effectively identify key stock nodes with greater risk influence from the stock network.