Correlation Research On Microblog Content Mining And Financial Time Series

With the development of Internet,social media has become part of people's daily life.Social media provide users with an instant sharing platform,which enable users to share messages via texts,pictures,audio and video,and the social property enable those messages to spread exponentially along the focus chain.Social media is not only the important message sharing platform,but also message acquiring source for netizens.Along with the rise of mobile internet,microblog has been developing rapidly due to its mobile features.80 percent netizens are using social network globally.In China,according to report of CNNIC,the population of China's netizens has reached 668 million by June 2015,and 70 percent of them are users of social network,30 percent are users of microblog.By the end of third quarter of 2015,monthly active users of Sina Microblog has reached 200 million,which makes Sina Microblog the most important social media in China.The prosperity and development of microblog brings huge amount of content information.Thus data mining on such user generated information has important significance.Content of microblog has great value.On one hand,those information reflects the attitude towards the users own life,on the other hand,it contains opinion and advice towards livelihood.It is very important to extract public opinion toward financial domain and events,and analyze its relation between the result of public microblog content mining and financial issues for personal and organizations' investment strategies making and regulatory agencies'policy making.For microblog content,there are quite a lot to mine out.To begin with,the hiding message beyond the microblog posts should be extracted.Then,the social network reflected along with the spreading chain should be extracted.Based on the above idea,this paper makes an attempt to conduct both sentiment analysis and graph analysis in order to extract the hiding information in the text and structure of microblog.A relation method is proposed in this paper to relate microblog to financial entities so as to apply microblog mining results.Correlation analysis is conducted based on the relation method.Then microblog mining results are used to help predict financial time series.And based on the prediction results,an automated trading strategy is applied to verify the efficiency of prediction of financial time series with microblog content mining results.