Statistical Studies Of Network Public Opinion’s Investor Sentiment’s Impact On The Stock Price

Social media-based blog, micro blog, online forums are filled with the entire Internet space at a rapid pace. In social media, everyone is not only a publisher of information, but also an audience. Internet evolved from a simple information technology platform to the main carrier of social media. In particular for China’s stock market, the development of the stock market in the past20years has been maintain high-speed. While the basic stock trading technology in sync with the international standards, the investment environment, stock trading system, and the effective regulation of the stock market are still exist certain defects. The market is non-validity. Social media information on the behavior of investors in the stock market is quite a mixed bag, which the false, negative and not yet been confirmed information spread quite fast. These information interact each other, and converging to a the powerful power, which plagued the healthy development of China’s stock market. Therefore, how to determine the impact of public opinion information on the stock price more accurately become of great significance.Based on the above considerations, the paper’s theme is set to the statistical study of the network public opinion’s impact on the stock price. First of all, focus on combing the existing research on the impact of the network public opinion on the stock price. and the lack of research is mainly reflected in two points:First, the theoretical foundation is weak. The second is the lack of information utilization. For the first question, the article sort out related research draw relevant research is quite mature. But the research considerate about the subjective factors is not enough. So the traditional theory can’t explain the fluctuations of the share price. At this time, behavioral finance theory happens to provides a theoretical basis for the related research. This article introduces FF three-factor model to compensate for the lack of existing studies theoretically weak. For the second question, the article use text information of the network public opinion to build the investor sentiment index. Then introduce investor sentiment into the model. The article draw the following main conclusions.First, when the stock price is in the decline time, the investors will pay high degree of attention to it.Second, when the stock market fluctuate a lot, the behavior of investors prone the same. And to some extent, the investor behavior exist "herding".Third, investor sentiment can predict the stock price to some extent.Fourth, investor sentiment and investor attention can explain the stock return volatility to some extent.Finally, start from the subjective point of view, when the network public opinion included in the analysis of fluctuations in stock prices, we find that it is a good complement to the traditional analysis methods.The possible innovations of the article are as follows:First is the research perspective. This article attempts to pick up the emotional information from the network public opinion. And analyze the impact of network public sentiment information on stock price. View to offer for the relevant authorities an example which is worth learning.Second are research methods. This paper innovatively use network public opinion data, and adopt text classification technology to build investor sentiment index.Third is the conclusion of the study. Compared with the conclusions of related works, this article draw a conclusion that network public opinion information can predict the stock price to some certain.The study was limited to Preliminary research. There is still more to be desired:First of all, the selection of research methods of this study is cross-cutting and multi-disciplinary themes. Research methods introduced this theme is still at a preliminary stage. And this is also an attempt based on the author’s thinking. Due to the limited capacity and inadequate preparation, this article is inevitably has some omissions.Second is the choice the measure method. The available of the investor sentiment measure are subjective method, objective method and composite method. This article choose subjective methods. And use text classification techniques to measure. Measure the emotional value is a systematic project, which takes a lot of manpower, material and financial resources. As my capacity constraints, and the associated workload is quite huge, the measure of investor sentiment in this study can only done by off-the-shelf sentiment classification software. Future study will try to develop their own software for the characteristics of its own research.Finally is the conclusions of the study. The conclusion is the network public opinion has an impact on the stock price. It is clear that this conclusion is not clear different with the existing study. After a preliminary analysis, it may be due to the limited sample size selected. And the most important reason is that text data mining and text classification in the actual operation cost a lot. If we grab the data of all the listed companies, it should use particularly large capacity memory storage. And this can not be achieved in the short term. Follow-up study can try to establish this database.