A Web Agent Based on Exploratory Event Mining in Social Media

We introduce an exploratory event mining system in social media. A system deals with user's interests by making use of interaction between user input and system output. The system presents related events in the form of graph structures called ``event graphs''. Event graphs based on users' interests are produced by iterating over event presentation and user selection. The system's user interface helps users obtain a better understanding of the content and background of news events. Our system has two kinds of agents. One is a news agent to detect related news articles and the other is a microblog agent to detect related microblog posts on the Web. A microblog agent detects related microblog posts using news articles that contain news URL references, and using a similarity between news titles and tweets.