Link Prediction and the Evolution of Communities on Twitter

摘要: This research is about the influence of link prediction on evolution communities Twitter. We collected tweets from three technology micro-bloggers who led us through their followings and to tens thousands unique users over several weeks. analyzed conventional alternative information streams for these based URLs embedded in followees followees-of-followees. model most recent latest they follow, which we infer links extract semantic entities that are indicative interests. Furthermore, propose a pipeline methods user modeling personalization interest test performance different organizational principles community design, including hierarchy, interests baseline follower mechanism Twitter, intuitions. The goal this thesis create better notion by automatically calculating adaptive personalized structures produce highly interesting content. Designing way use- ful because it enables people know organized during given period time community-based recommendations. designing automatic construction. Currently, manually constructed tedious process following unfollowing disconnected investigate whether possible between Twitter not explicitly connected explore such inferred social networks would allow improving content recommendations