Research On The Value Of UGC Knowledge Features And Sentiment Features To Stock Market

"Internet+finance" is an important direction of China’s economic development.By combining finance with big data,artificial intelligence and other new technologies,and exploring ways to change the information collection sources and investment decision-making process of traditional finance,the service efficiency of traditional finance will be greatly improved.In order to investigate the crowd wisdom of usergenerated content(UGC)and its influence,this thesis studies the financial influence of different UGC posting types on investment-oriented social platforms.User-generated content contains emotions.In the process of information transmission,investors are more likely to be influenced by other users’ emotions and thus change their investment decisions.So,it is necessary to study the knowledge and sentimental characteristics of user-generated content.In summary,the main research content of this paper is divided into three parts:(1)study the knowledge characteristics and market response of usergenerated content on investment-oriented social platforms;(2)study the emotional transmission process among different UGC posting types and its influence on stock price;(3)predict the behavior of stock price based on the time-series feature of stock price,user-generated content knowledge feature and sentiment feature.The main research steps of this thesis are as follows:(1)based on the background of strategic management theory,subject features are extracted from UGC through LDA subject model,subject features are evaluated based on risk-based knowledge,and features of UGC risk knowledge are further extracted;(2)based on the theoretical background of crowd wisdom and behavioral finance,this research extracted emotional features from UGC by means of lexicographic emotion analysis,constructed an econometric model of emotion-stock price,and conducted an empirical test on the transmission relationship between different UGC posting types of senses and their impact on stock price;(3)based on the efficient market hypothesis,by the wavelet frequency processing of share price through decomposition reconstruction of temporal characteristics of history price feature extraction,and then respectively with UGC knowledge characteristics,the UGC sentimental characteristics constitute a benchmark model,through the comparative study with the benchmark model,further put forward the fusion of three kinds of feature fusion model the accuracy of the prediction of stock price behavior level.In the context of the vigorous development of social media and fintech,this research can explore the tacit knowledge behind massive amounts of ordinary usergenerated content to a greater extent through the treatment of themes and emotions in the user-generated content,so as to improve the decision-makers’ cognition of the changing trend of the securities market.The conclusion of this paper is to predict the financial market by integrating UGC knowledge features,emotional features and stock price time series features.The result proves the existence of wisdom of intelligence The research content of this paper is helpful to observe the price fluctuation behavior of the stock of listed companies in the short and medium term,providing reference value for investment decision-making,as well as providing reference and warning function for market supervision and company supervision.