A Semantic Model for Social Recommender Systems

Social recommender systems, which have emerged in response to the problem of information overload, provide users with recommendations of items suited to their needs. To provide proper recommendations to users, social recommender systems require accurate models of characteristics, interests and needs for each user. In this paper, we introduce a new model capturing semantics of user-generated tags and propose a social recommender system that is incorporated with the semantics of the tags. Our approach first determines semantically similar items by utilizing semantic-oriented tags and secondly discovers semantically relevant items that are more likely to fit users’ needs. Experimental results show that our model offers the quality of recommendations in social recommender systems.