Research On The Construction And Application Of Investor Sentiment Index Under The Background Of Big Data

Traditional finance believes that investors are "rational people",which will be corrected immediately when the market price deviates from the value.Since the end of 1970 s,there have been many anomalies against the traditional financial theory in the financial market.In this regard,behavioral finance studies the psychological characteristics and decision-making behavior of investors in the process of market participation from the perspective of psychology,and forms a relatively complete theoretical system.An important research content of behavioral finance is investor sentiment,and the quantification of investor sentiment is the basis and focus of related research.Before the era of big data,scholars mainly selected some indicators that can reflect the overall development of the stock market,such as trading volume and turnover rate,as proxy variables of sentiment.Such variables can not fully reflect the changes of investor sentiment,and can only objectively reflect investor sentiment.With the advent of the era of big data,the features of Internet data,such as large amount,comprehensive,real-time and real-life,provide a more comprehensive data base for the compilation of investor sentiment index.At the same time,with the continuous maturity of machine learning and other technologies,the data that could not be obtained and used originally due to technical limitations are gradually included in the scope of research by scholars.Therefore,based on big data construction Investor sentiment index is a relatively objective and accurate measure of sentiment.This paper mainly studies the construction of investor sentiment index and its impact on China's stock market.As for the construction of investor sentiment index,firstly,based on the definition of investor sentiment from the perspective of big data,the index system is constructed by selecting "user comment content" which can reflect investor sentiment tendency and "user post quantity" which can reflect investor attention.Then,based on the data acquisition and preprocessing module,machine learning classification module,and investor sentiment index compilation module,this paper introduces the idea of investor sentiment index construction.Finally,based on the big data technology,this paper obtains the massive text data of Dongfang wealth online stock exchange index bar and preprocesses it.By constructing the LSTM machine learning model,it classifies the emotions contained in the user comment data,counts the numberof positive,negative and neutral emotions in the user comment data of each trading day,calculates the daily investor sentiment index and carries out the following work The effectiveness test provides basic work for the research of investor sentiment.As for the application of investor sentiment index,this paper constructs the vector autoregressive model of investor sentiment and stock market return and stock price volatility respectively,and studies the relationship between investor sentiment and stock market return and volatility.The empirical conclusion is that stock market return and investor sentiment are the Granger cause of each other,and the change of investor sentiment can significantly affect stock market return.The same stock The stock price volatility and investor sentiment are one-way Granger's reasons.The change of investor sentiment can significantly affect the stock market volatility,but the impact of stock market volatility on investor sentiment is not significant.The main innovations of this paper are as follows: 1.The LSTM machine learning classification model is constructed to mine the emotions contained in the massive user comment data.This model can provide a better solution to the problem of user comment classification with long content.2.In the formulation of investor sentiment index,not only the number of Posts expressing positive and negative emotions,but also the number of Posts expressing neutral emotions are considered.3.In the aspect of the application of investor sentiment index,different from the previous empirical research,which generalizes the stock market volatility and return volatility,the paper studies the relationship between investor sentiment and stock market return and volatility respectively.