Text Analysis And Data Mining Based Stock Analysis System

Users can only obtain real-time stock market data and independent information about the review of stocks between stocks from traditional stock trading software.It is difficult for users to find hotspots in the current market from such information,and such information is also difficult to associate with text data of stocks that users consider during stock trading such as news,announcements,etc.At the same time,this traditional way of analyzing stocks is basically based on the unconnected way of thinking between stocks.However stocks in the Chinese A stock market are not independent of each other.When price movement occurs,the changes of stock price show a kind of cluster behavior and those hehaviors have strong relation with the news information and stock announcement which is exist in external stock market.The purpose of this system is to analyze the stock correlation through data mining and other technologies to find the relationship between stocks.In turn,a more detailed analysis of the Chinese A stock market is carried out based on the relationship between stocks and market data.At the same time,text analysis technology is used to further correspond to text data and analysis results such as news and announcements.Assist the user finds the reason for the stock's rise and fall.This system mainly based on the similarity calculation,association analysis,clustering technology in data mining and text summarization and text generation in natural language processing.In order to solve the problem of isolation between stocks,the system first proposed the concept of tag,which is a fine-grained and hierarchical division of stocks.Then,establish the relationship between the stock and the tag respectively.Based on those relations,combined them with the stock market data,and using the ideas of multiple influencing factors,this paper provide analysis of the Chinese A stock market from the perspective of integrity and relevance.In order to test the correctness of the stock relationship and the reliability of the analysis results,this paper designed the corresponding experiment and verified the result.The experimental results show that the proposed method has certain accuracy and strong interpretability.In order to make a strong correlation between the analysis result and news information,stock announcement and other data outside the stock market.This paper make a sentence level document summary for news information and stock announcement to pick out key textual information associated with the analysis results.Finally,in order to better show the results to the user,In this paper,the analysis results are presented in the form of a network diagram,and the template filling based text-generated technology is used to describe the analysis results in a textual manner.This paper completed the construction of the entire system,so that the analysis work of this paper has better practicality.This paper apply the data mining and natural language processing technology to the analysis for Chinese stock market,it explored the feasibility of using those technology for stock analysis field.This paper proposes a stock market analysis method with integrity and relevance,and innovatively presents the influencing factors of the stock market as a closely connected network diagram.The work of this paper provides new ideas for stock market analysis.