The Sentiment Recognition Of Retail Investors Based On The Comments On The Stock Bar Forum

The stock market is a barometer of the economy,revealing that the law of stock market operation is of great significance for investors to make investment decisions.Factors affecting the stock market trend include: macroeconomics,industry development,internal factors and emotional factors of investors.The first three have changed less in the short term and it is difficult to reflect the real-time fluctuations in the stock market.The emotional factors of the stock market are subject to volatility due to factors such as news,stock market trends and stockholders' psychology.In economic activities,the information obtained by economic agents is often asymmetrical.Behaviors that are under the condition of information defects often distort the nature of behavioral equilibrium.That is,in the case of a large number of noise traders,the behavior of rational investors is difficult to return stock prices to basic values.Moreover,retail investors often have emotional cycles,and investors' own emotions have an important impact on the investment decisions made.In addition,retail investors as groups are also overreacting,that is,for a series of positive or bad news.Overreaction caused the stock market to rise or fall.The proportion of retail investors in China's A-shares is relatively high,and its volatility is largely influenced by the investment decisions of retail investors.Therefore,identifying and quantifying the sentiment of retail investors is of great significance for the healthy and stable development of China's stock market.In order to identify the sentiment of retail investors,this paper establishes a mixed model of investor sentiment recognition based on long and short memory model(LSTM)and naive Bayesian model(NBM),which divides the processed text into long sentences.With short sentences,the long sentences separated are selected for analysis using LSTM,while the short sentences are selected for analysis by NBM to form a hybrid model to improve the recognition accuracy of the model.This article uses a well-known stock bar 2018-9-27 to 2019-2-12 Shanghai Composite Index shares a total of 30 W comments,and use a mixed model to classify the emotional sentiment of the comments,generate a sentiment index,and the sentiment index and the Shanghai Composite Index Correlation test,VAR modeling and Granger causality test verify that the research results of this paper have certain predictive ability for the Shanghai Composite Index.