Research On The Interaction Between Investor Sentiment And Abnormal Volatility Of Stock Market Based On Stock Review Text Mining

Investor sentiment is one of the key objects to study the law of stock market movement.The traditional indicators to measure investor sentiment mainly include the direct indicators based on survey,the indirect indicators based on objective data from the stock market and the comprehensive indicators constructed by both of them.However,the traditional indicators have some limitations.With the development of the Internet,most investors comment on the Internet to express their opinions or emotions,which makes it possible to extract investor sentiment from online investor reviews by text mining.It provides a new perspective for in-depth analysis of investor sentiment and abnormal volatility of stock market.In recent years,scholars pay more attention to the relationship between investor sentiment and abnormal volatility of stock market.To interpret the abnormal volatility caused by investor's decision-making behavior from the perspective of sentiment will help to improve the existing system of behavioral finance and develop the theoretical content of behavioral finance.Based on this,this paper use text mining technology to quantify online investor reviews with a view to constructing a reasonable comprehensive investor sentiment,and then studies the interaction between investor sentiment and abnormal volatility of stock market.Focusing on the problem under study,firstly,get online text review data through web crawler.After a series of data processing,use LSTM algorithm to build sentiment classifier to classify text comments and quantify the online investor sentiment.Then,combine with indirect investor sentiment indicators to construct a new comprehensive investor sentiment through principal component analysis.Secondly,based on the theory of stock market feedback mechanism,analyze the relationship between investor sentiment and abnormal volatility of stock market.Use HP filter to identify the abnormal volatility of stock price and establish GARCH model and VAR model to analyze the clustering of stock price volatility and the interaction between investor sentiment and abnormal volatility of stock price.The empirical results show that investor sentiment and abnormal stock volatility are reciprocal Granger causalities and stock price volatility are significantly volatile,concentrated and highly persistent.Meanwhile,changes in investor sentiment will positively accelerate stock price volatility,and vice versa.Besides,the volatility of the stock market has a more intense stimulating effect on the volatility of investor sentiment.The innovation and contribution of this paper are as follows.First,using text mining technology to quantify text reviews enriches the comprehensive investor sentiment indicator.Second,study the interaction between investor sentiment and abnormal volatility of stock market in deepth.Simultaneously,verify and improve the theoretical content of behavioral finance from empirical perspective.At the same time,the results of the study are of practical significance for investors and the relevant government departments to understand the sentiment tendency and stock market movements.