Discovery And Analysis Of Important Nodes Of CSI 300 Stock Complex Network Based On LeaderRank

With the advent of the era of big data and the continuous deepening of research on the stock market,people realize that the stock market is a complex system.Using complex network methods to study the correlation and structural characteristics between stock markets,we can thoroughly understand the complex relationships and internal connections of the stock market as a whole.With the deepening of research,people have discovered that the correlation between stocks and the strength of influence in the stock market are not symmetric.Therefore,how to build a reasonably realistic stock network is the key to understanding the stock market.At the same time,there are some important stocks with strong correlation and high influence in the stock market.These important stocks can be accurately excavated and empirical analysis can be provided to provide relevant departments with effective suggestions for the management and control of stocks.In this paper,the Shanghai and Shenzhen 300 data is used as a research sample.The Granger causality test is used to characterize the causal relationship between two stocks.The optimal threshold method is used to denoise and simplify,and build a directed weighted stock network.Then divide the stock community into a complete and independent hierarchical structure.Leader Rank algorithm was used to mine and sort the important nodes of the stock network under the optimal threshold.The analysis found that the stocks of various industries have different positions in the stock market.The mining,manufacturing and financial industries are the three most influential industries in the stock market.Among the two major societies,there is a clear phenomenon of aggregation.If a stock in an association fluctuates abnormally,a linkage effect between the community and the community is likely to occur,and the risk is easy to spread within the community.Therefore,the relevant regulatory authorities shouldstrengthen the supervision and control of the important stocks of these large industry players to prevent risks Spread between stocks.The experimental results show that the directed weighted stock network constructed using Granger causality test is a scale-free network with a small power-law index of the degree distribution,indicating that the stock market has more hub nodes than other real networks There are many stocks that play an important role in this,which helps the stock market maintain balance and prevents the monopoly of several stocks.This paper adds a weight free adjustment parameter based on the standard Leader Rank algorithm.The mining of the important nodes of the directional weighted stock network constructed in this paper is more targeted,well adaptable,and has a fast convergence speed.It is more accurate to discover and mine.Important stocks in the stock market.