Research On The Influence Of Internet Public Opinion On Stock Return

The rapid development of Internet technology has caused a great impact on economic and social development and people's lives.The emergence of social media platforms has rapidly changed the way and path of information transmission.In the capital market,the relationship between investors and enterprises has become closer because of the Internet platform.The appearance of Weibo,stock bar comment area,and We Chat circle of friends has enabled the majority of small and medium investors to complete the role change from passive receivers of information to signal senders.Specifically,investors made their voices through comments on social media platforms,reposts,etc.,to express their emotions,forming a huge public opinion force.This investor sentiment has also chemically reacted with the capital market.In the stock market,fluctuations in investor sentiment may have a significant impact on the company's future stock price performance,stock expectations and actual returns.This paper selects the weekly stock return data of the top 70% of the company's A-share market capitalization from September 2018 to September 2019 as the research sample.The reason why the sample of 70% of the market capitalization was chosen is to consider that there may be more potential shell companies in the domestic capital market due to strong IPO supervision.Then use the web crawler technology to obtain the text of the investor's review of the sample company's stock in the corresponding period of time in the Oriental Fortune Stock Bar,and use the Python stasis library and the lexical text matching method to analyze the positive and negative sentiment of the text to construct an investor sentiment index TM,referring to the construction method of the BW index,selects seven indicators of trading volume,transaction value,turnover rate,first-day IPO volume,closed-end fund discount rate,consumer confidence index,and Chinese investor confidence index,combined with the TM index Using the principal component analysis method to construct the investor sentiment measure SENT of this paper.In the empirical analysis,this article first uses the Fama-French three-factor model to investigate whether there is a value effect and a scale effect in the domestic stock market.The specific method is to classify the sample according to the scale factor and the value factor and then cross-combine them to obtain four different investments.Combining the samples,the regression results show that there are significant value effects and scale effects in the domestic stock market.Considering that the domestic capital market is under the background of IPO and strong overseas investment supervision,the prediction accuracy of the Fama-French three-factor model may be reduced.Therefore,this article includes the CH-3 model,which is based on the Fama-French three-factor model.The model is aimed at the optimization of the domestic market,and in terms of scale factor,it excludes potential shell companies with a market value of the last 30%.The value factor selects the reciprocal of the price-earnings ratio rather than the company's book value as the proxy variable.The regression results found that the prediction accuracy based on the CH-3 model was higher than that of the Fama-French three-factor model,and the value effect and scale effect at this time still existed.Finally,this article breaks away from the framework of traditional financial asset pricing models and attempts to optimize the CH-3 model based on behavioral finance theory.The specific approach is to incorporate the constructed investor sentiment index SENT as a new factor into the model.In this paper,a model including four factors of market,scale,value,and sentiment was obtained,which is named CH-4 in this paper.Based on the regression results of the CH-4 model,it was found that investor sentiment has a significant positive impact on the company's stock return.After classifying the company,it was found that this effect mechanism only works in a small sample of companies.On the other hand,compared with the Fama-French three-factor model and the CH-3 model,the determination coefficient interval of the CH-4 model is improved to 0.77-0.85.This shows that the newly constructed model can predict the volatility of stock returns more accurately.The main conclusions of this study are summarized as follows: 1.Under the support of behavioral finance theory,it has been confirmed that investor sentiment has a substantial impact on company stock returns,and this effect only plays a role in small-scale companies.2.In view of the differences between the Chinese capital market and foreign capital markets in terms of regulatory strength and accounting indicators,the Fama-French three-factor model can predict stock return fluctuations,but the accuracy is not enough,and the improved CH-3model and the CH-4 model can better predict the stock market volatility.3.In China's stock market,there are also scale effects and value effects,which also shows that there are a large number of irrational investors in China's stock market participants,and their follow-up behavior will cause sharp fluctuations in stock returns.At the same time,the research in this article has certain theoretical and practical significance.