A Study On The Impact Of Self-media Sentiment On Stock Market

The development of Internet technology has given birth to new forms of information dissemination: the atomized mediums of scribbles,newspapers and journals have been replaced by graphics,videos and audios composed of bits,which are able to reach the consumers of the content more quickly and in a more timely and accurate manner due to their speed of dissemination and the amount of information transmitted without the constraints of the physical world.The Internet is composed of one node,which can communicate with each other.,This structure allows information dissemination in the Internet era to include all previous forms,such as point-to-point,single-point-to-multi-point,multi-point-to-multi-point,etc.Therefore,the users of each node can become media in the traditional sense,which is defined by Wired magazine as "everyone to everyone communication".In China,a typical example of new media in terms of graphics is the public number function on We Chat,which has 1 billion daily active users,where every individual with the desire to express can spread information by writing articles,a process also in line with the American scholar Lazarsfeld’s(2012)"two-level communication theory",will produce opinion leaders,and Corey(1971)argues that these specific people with strong influence will disseminate the information they receive or the information they generate and thus influence the decisionmaking behavior of others,Bollen et al.(2011)analyzed the daily text content on Twitter through different sentiment analysis tools to derive different sentiment states and found that changes in certain sentiment variables matched changes in the DJIA(Dow Jones Industrial Average).Antweiler and Frank(2004)hypothesized that people think about discussing their thoughts and posting content before trading broadly assuming that people trade stocks and that this content may lead to changes in trading volume and thus potentially predict volatility,ultimately concluding that this content in bullish sentiment has a predictive effect on volume and stock price.As a result,the topic of this paper is "A study on the impact of self-media sentiment on stock market ".Firstly,we crawl the content of articles and article traffic indicators of eligible financial public numbers,analyze the text for sentiment tendency through ERNIE natural language processing model,and then construct indicators that can respond to self media sentiment as explanatory variables,and obtain GEM composite return,trading volume from CSMAR database,The relationship between these two variables is explored.