Research On The Influence Of Public Opinion Index On Stock Price Trend

With the development of Internet technology,WeChat public accounts,video accounts,blogs and other new media have also experienced explosive development on the basis of the maturity of traditional media.It is said that in the era of information explosion,investors are often overwhelmed in the face of massive information.How to achieve efficient information acquisition and make up for the information gap between the market and investors has become a common pain point for investors.In recent years,the rapid development of quantitative trading technology has made it possible to find effective information through data mining to build investment strategies that achieve excess returns.Therefore,the reality urgently requires us to use information technology to process massive amounts of information and dig out public opinion indicators,in order to effectively predict the market and form quantitative trading strategies through this indicator.We first capture stock public opinion news published on online media,print media,Penguin,and WeChat public account platforms from December 1,2018 to October 31,2021,and uses sentiment analysis technology to map 1,885 stocks(Selection based on CSI 1000 constituent stocks and CSI 800 constituent stocks)in the A-share market.The news text is analyzed and scored to construct a public opinion score.Secondly,we constructed a random-effects panel regression model and a PVAR model to verify that there is a significant positive correlation between financial public opinion and stock price fluctuations in the short term through Stata.Positive public opinion orientation will promote stock prices to rise,while negative information will cause stock prices to fall.Moreover,the research results show that there is a significant causal relationship between financial public opinion and the rise and fall of stock prices,yet the influence of financial public opinion on stock prices has a significant timeliness,and its influence will reach a high point in a short time after the news is released.Finally,this report verifies the validity of the factors through the ICIR analysis method and the hierarchical backtesting method,which confirms that the factors have a certain ability to select stocks in the real quantitative trading market.To sum up,through the relatively complete coverage of the public opinion information on the whole network and the special adjustment of the public opinion thesaurus,the financial public opinion factor we constructed has achieved good effectiveness and stability,and has a certain independent stock selection ability.If we combine other factors to construct a multi-factor stock selection model,we can expect better returns.In addition,further optimization can still be made according to the characteristics of public opinion factors to mine public opinion factors to obtain higher alpha,such as further screening the quality of the public opinion corpus,improving the factor coverage,adjusting the warehouse adjustment cycle,etc.