Research On Investor Sentiment And Stock Market Based On Microblog Text Mining

With the coming of Web2.0 era, the various social networks (micro-blog, blog, forum, WeChat) play a more and more important role in people’s life. Web 2.0 lets people’s role change from passive recipients to active creator in information transformation, which change greatly the enthusiasm of Internet users. Everyday there are a large number of the original user data (User Generated Content) to produce. These data aggregation reveals popular views of social issues, such as the financial media, financial practitioners in social media, real-time publishing information in stock market, investors real-time access to this information, to express their views on the market and stocks, and the network participants discuss on certain problem and other social. These UGC data represent a certain period of time the market for the current and future period of time to judge, has important significance for predicting the trend of the stock market, more and more scholars to participate in the stock market the UGC data mining research. The existing studies consider information activity intensity and emotional tendency analysis of social network is conducive to the improvement of the existing stock forecasting technology, and the existing research at home and abroad to the forum and other traditional social media as a source of data for research object, micro-blog this emerging social media Co., based on the background of sina micro-blog, through information mining Sina micro-blog potential themes and emotional tendency, and explore its correlation with the stock market behavior, and to predict the stock market behavior related. The main work is as follows:(1) Investor sentiment. For the common investor sentiment analysis, it always has some indirectness. But this research uses the weibo content to analyze investor’s sentiment which are the true emotional express of these investors. Contrast to the indirect analysis, this research is more valuable to investor sentiment(2) Build investors sentiment index system. Based on previous research, this paper build the investors agent sentiment index system which lay a foundation of the future research.(3) Experimental design and research. Firstly, this paper tries to obtain the keywords from the stock market and sentiment dimension. Then, this paper tries to quantize these indexes. Through the empirical research, this paper finds that there are significant correlation between investor’s emotional tendencies and stock market indicators, investor’s expectation and stock market indexes, investor’s attention and stock market indicators. The model with investor’s emotion is much more precise than traditional model.