Study On Stock Correlation Networks Based On Artificial Stock Market With Diversified Information

The stock market is an important financing place for modern economic system.It is also a benchmark for the operating conditions of a listed company and an investment channel for investors to maintain their value of assets.The dynamics of a listed company‘s stock price are often determined not only by its operating conditions,but also by factors such as macroeconomic information,industry outlook information,investor sentiment,and by even the rating agency‘s assessment of its value.In addition to the dynamics of individual stock prices,various types of information also affect the correlation between stocks.For example,macroeconomic information and sector information have an impact on the rise and fall of the entire market and stocks belong to the same sector.Investor sentiment may lead to correlations between stocks that caused by herding behavior.Due to the extremely complex information environment in the stock market and the limited rational and irrational trading behavior of investors,there are various complex correlation features such as lead-lag,price synchronization,industry rotation,and risk contagion among stocks.These correlations form a network consisted of stocks.If the rise and fall of stocks no longer corresponds to the intrinsic value of listed companies,then the efficiency of the stock market declines.This not only limits the function of financing and allocation of the stock market,but also directly affects the investor‘s investment objective of diversifying risks and gaining returns.In view of these,it is necessary to study the stock correlation and stock correlation network caused by various types of information in order to achieve an in-depth understanding of the intrinsic correlation between stocks,and to monitor the risk contagion among stocks.For this reason,this paper builds a diversified information artificial stock market based on the existing research,starting from the various characteristics of diversified information to study the causes of correlation between stocks,and the influence of diversified information on the characteristics of stock correlation networks.First,a diversified information artificial stock market model is constructed,and on the basis of describing the asset structure in the stock market,the dynamic characteristics of investor wealth evolution are explained.After comparing the advantages and disadvantages of various price formations and clearing mechanisms in the existing literature,the price formation and clearing mechanism of this article is determined.Then on the premise of investor heterogeneity,risk wealth allocation strategies under each belief of investors are discussed.Further,the article defines the diversified information and based on which processes of updating the heterogeneous beliefs using diversified information are analyzed..Second,based on the diversified information artificial stock market model,the causes of the correlation between stocks are simulated from the perspective of multi-information.These perspectives include: market fundamental information,sector fundamental information,individual stock fundamental information,stock intrinsic value information,investor sentiment information,and so on.The main research results show that the correlation structure of stocks‘ fundamental information is an important prerequisite for the formation of stock price correlation.The higher the accuracy of the fundamental information of the stock fundamentals obtained by investors through learning,the smaller the correlation between stocks.As the accuracy of intrinsic value information and the decrease in the number of insiders increase the correlation between stocks,the underlying mechanisms are ?public information driven effect? and ?limited wealth constraints effect? when investors risk wealth is distributed among different stocks.As market sentiment increases in investors‘ sentiment,the correlation between stocks increases and its formation mechanism includes both ?public information driven effect? and ?limited wealth constraint effect?.Finally,this article expands the number of stocks and stock types in the fourth chapter,and constructs stock fundamentals including market fundamental information risk factors,sector fundamental information risk factors and individual stock fundamental information risk factors to simulate real-world stocks: the ?market-sector-stock? hierarchy in the market.Starting from various characteristics of diversified information,the characteristics of the stock network,such as the strength and structure of the global coupled network,the characteristics of industry-based aggregation of stock network,the topology structure of the stock network,and strength changes in stock network are studied in-depth.The research results show that the increase of the fundamental information intensity and accuracy of market risk factors and industry risk factors all lead to an increase in the overall correlation between stocks,but the internal structural trends are different.Changes in the fundamental information intensity of market risk factors and industry risk factors all lead to changes in the topological structure of the stock network;information accuracy is only related to the strength of the association network.The increase of the fundamental information intensity of market risk factors and industry risk factors respectively weakens and enhances the industry clustering characteristics of stock related networks,and the improvement of information accuracy does not lead to changes in the industry clustering characteristics.The decrease in the accuracy of the stocks‘ intrinsic value information and the increase in the number of insiders lead to a decline in the overall correlation between stocks.The accuracy of the intrinsic value of stocks,the weight of market sentiment of investors,and the strength of the stock industry clustering are inversely proportional to each other;and the number of stocks with internal value closeness and the weight of sector sentiment information are proportional to it.Intrinsic value accuracy reduces the strength of the network and does not change its topology;the number of intrinsic value insiders is related to the topological structure and strength of the stock network.The overall correlation between the stocks and their structural changes tend to be quite different when the market sentiment and the sector sentiment increase in the investor sentiment.Both market sentiment and sector sentiment can simultaneously change the topological structure and correlation strength of the stock network.When the number of stocks is expanded and the ?market-sector-stock? hierarchy is divided by fundamental information,the correlation between stocks becomes more complicated and the trend of change is more diversified than the situation when the number of stocks is smaller.In short,based on the diversified information artificial stock market,this paper conducts an in-depth research on the causes of stock correlation formation and stock network from the perspective of diversified information such as fundamental information,intrinsic value information,and emotional information.It provides different theoretical perspectives for the understanding of the complex relationship between stocks,the sector clustering characteristics of stock networks,the topology and strength changes of stock networks.It also provides stock market investor and concerned regulators with some practical references.