Design And Implementation Of A Microblog Based Burst Prediction System

With the rapid development of Internet technology, social network applications arereccived many attenions because of their real-time and high coverage. In order to usethis vast amount of information effectively, more and more researches focus on theinformation extraction and analysis. However, there are two problems on the socialnetwork information extraction and analysis. First, the massive and scatteredinformation cannot be retrieved efficiently; second, the current researches only analyzethe history and current information, they usually don t use this information to predictthe future.In view of the above problems and in order to explore popular words or topics onthe social network during a period of time in the future, this paper proposed a pop-wordprediction system based on micro-blog. Since Twitter is a prevalent micro-blogapplication, this paper uses Twitter to sample and analyze information. Our system canfigure out the events and words, which might burst in the social networks in the futureaccording to the text information, and it could also analyze the behavior relationshipbetween user information and events and words. Our system also could calculate thetrend of spreading of these words and events. Some practical experiments have shownthe ffectiveness and accuracy of our system.