Research On The Relations Between Social Data And Stock Price Volatility

With the rapid development of Internet financial, more and more investors shift their attention to the platform of Internet social. Not only it can be more convenient to provide timely market information for investors, provide trading channels, facilitate investors to trade stocks, but also it has became a social sharing platform of financial information when it provides these services. When massive and immediate social data presented in front of investors, we can raise some questions about that Is it possible to predict the volatility of stock price in a degree?What social factors is more significant to the volatility of stock price? Can I help investors improve their investment strategy on the basis of the prediction model? This is the main content of this paper.The main objects of this research includes snowball network, Sina Finance, microblogging, Eastern wealth network and other Internet financial platform. Firstly,I find the main attention of research status in domestic is on network public opinion instead of other user’s shared data on the social platform.And I will find the more important social data to the stock price,and put forward the research hypothesis.Then, according to the study hypothesis to crawling relevant social data.As the same time I will make structured treatment for non-text data, and establish emotional text data corpus,to identify emotional features. Then select the appropriate intelligent algorithm to construct feature models to verify hypotheses, and build predictive models in a certain precision to predict fluctuations of the stock price.The results showed that there are some social factors like the new focus degree of a stock is more significant than the network public opinion. The reason for the existence of the influence of thesis factors is the real-time social data reflects the attitude of investors for the market development during that period of time.Then build the prediction model through Boosting algorithm by combining those social factors and the network public opinion. The research discover that the collected data of the day is more significant in the next trading day than other day, and showed a decreasing trend as time goes on. This article provides a contribution for social data and stock price volatility relationship to a certain extent, it enriched the theory of the field, and provide new ideas and methods for the study of the field. In addition, the research provide valuable predictive model for investors as reference.