Research On Financial Media News Sentiment And Its Impact On Stock Market

Asset Portfolio theory is an important cornerstone of classical finance.The two important hypotheses of this theory are investor rationality and efficient market hypothesis.Asset portfolio theory and subsequently capital asset pricing model,arbitrage pricing theory and option pricing model constitute the classical financial theory system.With the rapid development of capital markets worldwide,the increasing number of financial derivatives and the rapid growth of financial market traders,efficient market hypothesis has been challenged both theoretically and empirically and it is difficult to explain the anomalies in financial markets.Behavioral finance,which integrates behavioral science theories psychology into finance,is a new frontier discipline.It explains,studies and predicts the development and changes of financial markets through micro-individual behavior and the psychological motivations.The study of financial market sentiment is an important field of behavioral finance.With the increase of the amount and availability of financial information,the study of financial market sentiment has attracted widespread attention in academia driven by computer technology and big data technology.The study of financial market sentiment has attracted widespread attention in academia.A large number of academic achievements represented by investor emotion,emotion index,news sentiment emerges.The thesis systematically reviews the relevant literature on news sentiment research.Text Database and Emotion Database are used as research databases.It studies the relation between news sentiment and financial marktet from four aspects:investor equilibrium price model,news text sentiment research,individual stock news sentiment research and index news sentiment research.the main research contents are as follows:Investor Equilibrium Price Model: On the basis of discussing the definition,development and research topics of Behavioral Finance,the thesis constructs investor homogeneity equilibrium price model and investor heterogeneity equilibrium price model from the perspective of investor heterogeneity.It also deduces the equilibrium price under the conditions of investor homogeneity and heterogeneity and discusses news The effect of emotion on equilibrium price.News Text Emotion Research: On the basis of comparing various dictionaries of emotional words,this paper studies the quantitative method of financial news emotions based on "word bag".Loughran and McDonald’s financial dictionaries are used to extract emotional words of financial news and the proportion of positive and negative emotional words to news is used to quantify positive and negative emotions of news.Machine learning methods including naive Bayesian,support vector machine and neural network are used to identify the positive and negative of news,and the accuracy of various machine learning algorithms in news emotion classification is compared.Research on News Emotions of Individual Stocks: On the basis of quantifying financial news emotions using Loughran and McDonald’s Financial Dictionary,this paper studies the effects of news sentiments on individual stock returns and weekend sentiment accumulation in Text Database and compares the differences of news sentiments effect on Chinese and American stock markets.By studying the influence of announcements on the time distribution of media news reports,the thesis finds that time distribution of news reports has a strong effect of announcement aggregation and announcement emotional reinforcement.This paper studies the effect of news sentiment on stock volatility and constructs stock trading strategies based on news sentiment.The empirical results show that stock trading strategies based on news sentiment have stable returns.Research on News Emotion Index: This paper studies the method of constructing the monthly market sentiment index based on News Database.The first momentum emotion index and the second momentum emotion index are constructed by using the uncertain method.The empirical results show that the emotion index has a strong ability to predict the return of SP500 index.On the basis of verifying the correlation between mutual fund returns and news sentiment,the thesis constructs a mutual fund portfolio based on news sentiment index and demonstrates that the portfolio has stable returns.This study shows that financial market sentiment can affect stock returns,volatility and index returns by influencing investor sentiment and trading behavior.As there are more retail investors in China’s immature stock market,they are more likely to be disturbed by news sentiment.Therefore,the relevant regulators need to pay close attention to the changes of market sentiment and take timely measures to prevent the sharp rise and fall of the market caused by sentiment.