Empirical Analysis Of The Influential Nodes In Chinese Stock Market Based On Complex Network

In the process of the development of the stock market,it is influenced by many factors such as political factors,industry factors,natural factors and so on.In stock market,the fluctuations of stock prices are highly inter-coupled with strong correla-tions,and most of the influence of the factor are reflected by the volatility of the stock market.In fact,the fluctuations of the stock price create an evolving complex system,the relationship between individuals in the stock market is close and complex,some im-portant stocks usually play crucial roles in a stock market.It is of great significance to study the relationship and structural characteristics of the stock market and to deepen the understanding of the stock market.As the development of network science,the work of using complex network theory to study the complexity of the stock market has gained a lot of valuable results.People gradually realize that it is clear that the effects of different stocks are directional and asymmetric,in the real stock complex networks,an increasing number of the scholars focus on to establish the directed stock networks.At the same time,as we all know that different stocks play an important role in various levels,the greater the influence of stock play,the more important the node location,so how to effectively describe the stock node location,and dig out the important stock analysis has important practical value.In the paper,therefore,based on the data of Shanghai and Shenzhen A share,applying the Granger Causality Test to describe the correlation between stocks,a weighted and directed stock complex network is built up,and the influential nodes for stock complex network are ranked by using LeaderRank algorithm.Further,in order to better understand the structural change process of the stock network over time,we use the method of the rolling window and the minimum spanning tree(MST)to construct a dynamic and directed weighted stock network,and analysis the trend of the change of network structure and the proportion of the industry in which the influential stocks are belonged.The details are summarized as follows:In Chapter One,we firstly expound some problems in the process of the rapid development of the stock market,and then leads to the research goal and the signifi-cance of the research.The development of complex network and the research status of important nodes are reviewed and evaluated.Finally,the main work and framework of the paper are presented.In Chapter two,we firstly introduce the statistical method of building stock net-work,we next introduces the knowledge related to graph theory,then gives the defi-nition of network static characteristics in complex network,and finally introduces the algorithm used to excavate important nodes in stock network.In Chapter three,we aim to rank the influential nodes for Shanghai and Shenzhen A shares market by using complex network analysis approach.To begin with,making full use of the Shanghai and Shenzhen stock markets' daily closing prices for ten-year period from the year 2006 to 2016,applying the Granger Causality Test to describe the correlation between stocks,a weighted and directed stock complex network is built up.Secondly,the stock complex network is further simplified by the method of special threshold.Thirdly,the influential nodes for stock complex network are ranked by using Leader-Rank algorithm.Last but not the least,the empirical results show that the manufacturing industry is the most influential industry in China;many crucial stock nodes are mostly concentrated in the same community and have agglomeration phenomenon obviously.In Chapter four,we study the network characteristics of the Shanghai and Shen-zhen A stock market dynamic network and the trend of the important nodes ' trends during the period from 2006 to 2017.Firstly,the daily closing price data of Shanghai and Shenzhen A shares from 2006 to 2017 are divided into 11 time windows by using the method of time window.Secondly,by using the method of Grainger causality test,a series of stock networks axe constructed for the stock data in each time window,and by using the method of the minimum spanning tree to simplify the stock network,we get 11 different MST stock network,and analyzes the characteristics of the stock.Then,the important nodes of each time window are excavated and analyzed by using LeaderRank algorithm.The results show that:(1)the dynamic stock network is a scale-free network,the average distance of the network will be significantly increased before the financial crisis,early warning effect to the crisis;(2)both in Shanghai A stock market and Shenzhen A stock market,manufacturing show in the stock market position is increasing,but the financial industry and information technology of the two industries is different,the financial industry in Shanghai A stock market is more and more important,and the importance of the information technology industry in Shenzhen A shares is increasing.