Study On Agglomeration Effect Mining And Behavior Characteristics In Chinese Stock Market

China's stock market has developed rapidly,but from the perspective of investor structure and market performance,the market maturity is still not enough,and it is even more unpredictable than the developed country market.In order to better understand the Chinese stock market,this paper uses the daily return data of the CSI 300 Index constituent stocks for the eight years from 2010 to 2017 as the research object.As the overall representative of the Chinese stock market,the stocks are clustered and analyzed to help identify the clustering characteristics of their common movements.It is generally believed that the stocks in the same industry sector have more obvious trends of the same rise and fall,and are more likely to cluster.However,because the stock behavior is affected by many factors,other properties of the stock itself will also trigger a clustering effect.The influence of these factors will cause the performance of the clustering from the industry to be masked,diluted,and even distorted.Therefore,on the basis of previous studies,this paper uses a robust regression method to eliminate the interference factors outside the industry category,and effectively obtains more obvious indicators of common stock movements in the industry.On this basis,using the price trend removed of interference factors to capture the industry clustering effect and study the rotational relationship between indicators.The study finds that there are significant clustering phenomena in the Chinese stock market under the category of industry;in the stock market bubble phase,speculation and increasing leverage are not global behaviors,and the cluster purity scores of various industries have risen significantly during the stock crisis,and the overall pace is the same;as a hot spot for investment In 2015,the phenomenon of industry clustering of "real estate industry" after the stock market crash continued to increase,and the clustering phenomenon of relevant industries was promoted;the trend of clustering in various industries was strongly correlated between 2011 and 2017,and there was a significant industry rotation phenomenon.