Decision Support System For Investment And Regulatory Based On Internet News Text Mining

Securities markets (including the Stock market, Futures market, Gold market, the Bond market, the Fund market and Derivatives market, we mainly regard the Stock market as the research object in the paper) plays an important role on supporting the development of the national economy. An important place as a direct financing of listed companies, a growing number of economists and business people are concerned about China’s Stock market, which is closely related to the development of stock market and the national economy.In this case, it has also become a topic of great significance to explore the factors that affect the stock market volatility.The stock market is a very complex nonlinear system, and many factors affect the volatility of the stock price. In the theories which attempt to explain the impact of price changes, the theory of "information flow" gets broad support. The theory of information follows that flow of information drives stock trading volume and volatility. This theory has also been a mature foreign markets empirical research support. With the continuous development of network technology, the Internet has gradually become the main channel for public access to information, the impact of Internet information on the stock market has also become a keen research directions of scholars. A large number of domestic and international studies have shown that the impact of stock price volatility of Internet information. However, with a broad array of Internet information, how to let investors and managers more intuitive, convenient access to and use information, interpretation of the "behind the story" of Internet news and information, provide decision support, it will be a very meaningful research.Based on News quantitative indicators of stock returns prediction model, we will develop the "Decision Support System for Investment and Regulatory Based on Internet News Text Mining", the target users of the system are investors and managers of listed companies and securities markets regulators, through each module function of the system, they can get needed decision support. The main contributions of this paper are as follows:(1) cross-site the Internet news Scratch. Current financial news website, such as:financial news Web sites have different encoding formats and News, Sina, Yahoo, News caught climb development difficulties, financial news website coding, development Scratch, heavy workload and poor compatibility. Learn foreign the Web crawler study theory, self-development of the Internet news Scratch is parsed financial news website News, no need to encode for different sites, strong compatibility, high accuracy data of this study provide a strong supported.(2) News automatic text categorization. Text classification pre mainly rely on professionals pure manual, time-consuming, high cost, low efficiency defects. This article using automatic text classification technology, the Internet news text classification model training, in accordance with the good news category has been set, automatic news text divided into similar categories to facilitate the comparative study of the different categories of news on stock price volatility situation, but also for the managers of the securities market regulators and listed companies to provide decision-support utility.(3) Internet image index of listed companies. Traditional corporate image rating method mainly considers two aspects of the enterprise hard power and soft power. This paper studies major departure from the news text information perspective, abandon the traditional corporate image rating methodologies, by emotional word news text message analysis, and calculated in accordance with the self-designed algorithm, the Internet image index of listed companies, both user investment auxiliary decision making, can also be effective in helping the securities market regulator to regulate the securities market order to protect the securities markets fair and reasonable to run.(4) The development of the prototype system. In this paper, the Java EE technology is used to develop a complete set of regulatory and investment decision-support system, the system of current first quantified the relationship described in news text information and stock price volatility. System based on Struts2framework, combined with ExtJS interface design and the Mysql background database, good news and information on the listed company’s share price fluctuations intuitive image display, for each type of user (securities market regulators, managers of listed companies and investors) decision-making reference.Chapters of this paper are organized as follows:The first Chapter is Introduction. This chapter briefly introduces the background, significance, the problems which will be solved and a thorough review of research at both home and abroad. Then we point out that the main innovations of this paper and the main research methods and ideas.The second chapter introduces the Internet news crawler. We introduce the theory of news crawler, and then describe the news crawler the article uses in detail. The news crawler in this article can grab news from different websites.The third chapter is text mining technique. This chapter introduces the connotation of text mining and illustrates the process of text mining.The fourth chapter is the research about text classification of Internet-based financial news. This chapter introduces the principle of text classification technique, including the key techniques and key classification methods. Then, we classify the Internet-based financial news into categories set in advance by the methods of text classification technique.The fifth chapter is Internet image index of listed companies. We firstly introduce the basic knowledge of sentiment analysis and then do a research on Internet image index of listed companies, at last describe the process of calculating image of the index.The sixth chapter is the development of prototype system.The seventh chapter is summary and outlook. We summarize this major work done, and give out the future research directions.Through this system, it can directly and vividly show the information behind Internet news, and reveal their possible impact of stock price volatility, and provide decision support for various types of users.