The success and popularity of social network systems, suchas del.icio.us, Facebook, MySpace, and YouTube, have generatedmany interesting and challenging problems to the researchcommunity. Among others, discovering social interestsshared by groups of users is very important because ithelps to connect people with common interests and encouragespeople to contribute and share more contents. Themain challenge to solving this problem comes from the difficultyof detecting and representing the interest of the users.The existing approaches are all based on the online connectionsof users and so unable to identify the common interestof users who have no online connections.In this paper, we propose a novel social interest discoveryapproach based on user-generated tags. Our approachis motivated by the key observation that in a social network,human users tend to use descriptive tags to annotatethe contents that they are interested in. Our analysis ona large amount of real-world traces reveals that in general,user-generated tags are consistent with the web content theyare attached to, while more concise and closer to the understandingand judgments of human users about the content.Thus, patterns of frequent co-occurrences of user tags canbe used to characterize and capture topics of user interests.We have developed an Internet Social Interest Discovery system,ISID, to discover the common user interests and clusterusers and their saved URLs by different interest topics. Ourevaluation shows that ISID can effectively cluster similardocuments by interest topics and discover user communitieswith common interests no matter if they have any onlineconnections.
