Utilizing user tag-based interests in recommender systems for social resource sharing websites

摘要: Tag frequency, recency, and duration were combined to model the personalized preference.The social network was utilized find similar users in collaborative filtering.The incorporated system applied resource sharing systems. Recently tagging, also known as "folksonomy" Web 2.0, allows collaboratively create manage tags classify categorize dynamic content for searching sharing. A user's interest resources usually changes with time such a information rich environment. Additionally, is one innovative characteristic websites. The from provides an inference of certain interests based on this neighbors.To handle problem changing gradually time, utilize benefit network, study models user interest, incorporating tag-based information, performs recommendations using proposed method includes finding neighbors "social friends" by filtering recommending items content-based filtering.This examines system's performance experimental dataset collected bookmarking website. results show that hybridization preferences plays important role, better performances than traditional recommendation reveal friend can successfully collaborate, thus improving process.