The Impact Research Of Investor Sentiment And Stock Market Returns Based On Well-known Financial Micro-blog

In era of 4G,the Internet has accelerated the rapid development of social media.Information acquisition of internet users has gradually changed from passive acceptance to active creation,resulting in a large number of original information,including some generated by professionals.Sina micro-blog,as one of the important platforms,is selected as the data source due to its characteristics of wide audience and fast information dissemination.With financial column of Sina micro-blog defined as the target label,314 top micro-blog VIP accounts are selected and their data are captured by the crawler program.Nearly 3.78 million micro-blog text information from 2015 to 2019 is obtained and divided into three major categories: professional investment,forum and media.Through the text processing and emotional analysis,the integrated micro-blog collection data is divided into three major emotional categories: positive,neutral and negative,upon which investor sentiment indicators are built.This dissertation explores the correlation between the investor sentiment indicators and performance data of market returns,which is the combination of returns of Shanghai Stock Exchange,Shenzhen Stock Exchange,small and medium-sized enterprises board and Growth Enterprise Board.Four markets and three categories are analyzed by multiple linear regression model.The results show that:(1)Investor sentiment of micro-blog has a significant impact on the stock market,positive emotion positively correlated with earnings,and negative emotion negatively correlated with earnings;(2)There are disparities in the influence of investor sentiment of micro-blog on different markets.The influence of positive emotion gradually increases with increase of speculative attributes of the market,while the influence of negative emotion gradually weakens with increase of speculative attributes;(3)In the same market,there are differences between the emotional influence of different micro-blog investor categories.Analysis of the three categories of micro-blog data by regression shows that the investor sentiment of top micro-blog celebrities with professionalism and authority have greater influence.