| Chinese stock market has 33-year history of development,and the number of domestic investors officially exceeded the 200 million marks in 2022.As the quality of China’s economic development steadily improves and people’s lives continue to improve,the public’s demand for wealth management grows by the day.The Chinese stock market also has obvious emotional characteristics,and everyone likes to focus on hot stock targets,so how to identify the network structure and behavioral characteristics of investors in the stock market has become a topic of great interest to everyone.In this paper,we concentrate on the impact of network structure on the propagation of investor behavior,examine in-depth the process of sentiment transformation among various investor groups,and attempt to identify the sentiment intensity indicators to help us direct our investment behavior.First,this paper looks into the process of investor behavior propagation in the stock market and how network structure affects that process.Based on the epidemic propagation model and the dynamics of price formation causes,the investor behavior propagation model is defined to express the buying,selling,and holding behaviors of investors through the change of node states in the network.The buying and selling behaviors of investors infect each other in the network causing changes in supply and demand and expressed in price changes.The recovery of an infected node to a susceptible state is affected by the density of susceptible nodes and the recovery strength,rather than just the constant recovery time.The results of the simulation indicate that there is a correlation between the changes in the number of infected individuals in the investor behavior model and fluctuations in price.Additionally,the effective propagation rate exhibits a positive relationship with the return rate.Further analysis and comparison of the network structure among investors reveals that the model can accurately fit across different real data network,and the nearest neighbor coupled network with hub nodes is the best fit to the real network situation.This is consistent with the real situation of a small number of investors with huge investment funds versus many ordinary investors.Second,how investor sentiment changes and the factors that influence it in the process are investigated.This is because sentiment is often the main factor that drives investors’ behavioral decisions.By modeling and analyzing the changes in investor sentiment and the rules of subject switching in the investment process,the calculation of the main sentiment values and switching rules were improved based on the existing model of sentiment contagion,including sentiment types,contagion scenarios,switching conditions and other elements.The experimental results show that for the existence of different groups of investors on the network,although their emotion distribution,population proportion and activity patterns are different,eventually their emotions and densities will converge to a stable value with the development of time.And the intensity of the spread of emotions plays a dominant role in emotional infection.The amplitude of the individual’s emotion,the number of individuals following the transmitter’s emotion and the distribution of the individual’s emotion when it finally stabilizes are determined by the intensity of the transmission.Finally,applying the model to the real stock data,it is found that the propagation intensity of market sentiment is accurately detected and can be used as an accurate signal to predict future sentiment trends. |
