Personalized Recommender System on Whom to Follow in Twitter

摘要: Recommender systems have been widely used in social network sites. In this paper, we propose a novel approach to recommend new followees Twitter users by learning their historic friends-adding patterns. Based on user's past graph and her interactions with other users, scores based some of the commonly recommendation strategies are calculated passed into machine along recently added list user. Learning rank algorithm then identifies best combination user adopted add past. Although may not adopt any explicitly, they subconsciously or implicitly use some. If actually match ones suggested strategy, consider using that strategy. The experiment real data collected from proves effectiveness proposed approach.