Identifying Influential Taggers in Trust-Aware Recommender Systems

Trust-ware recommender systems provide the features of personalized product and service recommendations in web based social networks by using the trust connections existing between users and preferences data available for each user. One of the main sources of user preferences data are the tags that users apply to different items. Encouraging users to apply more tags is one of the challenges faced by most social network sites. In this paper we purpose an approach to identify influential taggers in a trust based social network so that efforts to encourage tagging can be achieved by designing incentives for motivating the influential taggers to apply more tags. In our proposed approach, for every user his tagging influencer is that user in his personal network who has influenced his tagging behavior the most. We define an active user tagging actions has been influenced by a user in his personal network only when the active user tags an item after his influencer has tagged it. The influential taggers in the overall social network are those who have the influenced the maximum number of users in the network. We analyze the real life dataset of Last.fm to show that our approach is different from the current approach of defining those users who have tagged the maximum number of items as the influential users. We also discuss the implications of using our approach.