Inferring the Interesting Tweets in Your Network

As the demand for quick, live and relevant information increases, more people look to microblogging sites, such as Twitter, as a source of content. Retweeting acts as a filter of useful information for users with more interesting information likely to be disseminated further through the network. "Interestingness" denotes the level of interest in a particular Tweet and we believe this has an influence on the retweetability of a Tweet. In this paper we introduce a method based on a Bayesian Network for inferring the relative interestingness of a Tweet based on its retweet history, including a scoring system for determining the level of interestingness. We show the results of our work in inferring which Tweets are interesting and validate the success of our scoring system in detecting globally interesting information.