Design And Implementation Of Stock Forecasting System Based On Text Sentiment Analysis

With the rapid development of Internet Finance and securities market,the number of new investors in China has been high,and stock bar and various forums have become an important medium for stock investors to share information and make investment reference.The theory of market ineffectiveness shows that the stock price can not fully reflect the stock value,that is,there is a certain correlation between investor sentiment and stock price trend.Based on the emotional analysis of stock reviews and the time series analysis of stock price related information of investors in the stock bar,we can predict the stock price to some extent,so as to provide investment suggestions for small and medium-sized investors.At the same time,the conclusion of emotional analysis can also provide the analysis basis for institutional investors and stock analysts.The factors used in traditional quantitative investment are generally stock related data or macro and micro economic data,and the market inefficiency is not modeled.This paper designs and implements a stock forecasting system based on text emotion analysis.The system can analyze the emotional tendency of investors on stock review in the stock bar,and then combine the analysis conclusion and stock price information to forecast and display the stock price.This paper uses the method of combining custom dictionary and deep learning to analyze the emotion of financial text,which improves the accuracy of emotional analysis of financial text.In the later part of stock price prediction,this paper proposes a combination forecasting model of ARIMA-LSTM.The model can make good use of the linear part and nonlinear part of stock price time series,combine emotional analysis conclusion and other custom characteristics to predict stock price rise and fall.The experimental results show that the performance of the combination model in prediction accuracy and error is improved,and toe model can complete the task of stock rise and fall prediction,and get excess return in the simulation disk test.This paper first analyzes and summarizes the related technologies and research background of text emotion analysis and stock prediction,and then models and verifies the effect of financial text emotion analysis and stock forecasting technology.Finally,the general process of software engineering is used to analyze the system demand,system design,module division,system implementation and system test.The final test and display results show that the system can complete the function of stock price and fall prediction and also meet the expectations in terms of stability and usability.