Identification of the most influential stocks in financial networks

In the financial market, the stock association network can be used to describe the correlations between stock prices. The rise and fall of stock prices and their mutual influence can be described by a reversible cascading failure process. In the reversible spreading processes, the influence of fluctuation of an individual stock price on companies, industries and even the financial markets can be regarded as the influence of the individual stock in the financial networks. In this paper, based on the stock price association networks of nine sectors in China A-stock market, we studied the identification of the most influential nodes in the reversible spreading processes. It is found that due to the high proportion of core nodes in the stock association networks, the k-coreness centrality can not accurately measure the influence of these nodes. A large number of simulations show that the number of out-leaving edges of neighbor nodes can better evaluate the influence of a node in the spreading processes. Then a node strength centrality and s-coreness centrality based on link importance are developed to measure the spreading influence of nodes. Compared with degree centrality and k-coreness centrality, the node strength and s-coreness based on strength have a better performance in identifying the most influential nodes. The ranking algorithm considering the structures and dynamics of local neighbors can further improve the ranking performance. The proposed framework and ranking method open up a new idea in identifying the most influential stocks in financial networks.