Ranking Tags and Users for Content-Based Item Recommendation Using Folksonomy

Selecting tags for describing the given item is the key to develop practical content-based item recommender systems using folksonomy. A novel strategy is proposed in the paper. With the strategy, tags are selected on the basis of users’ behavior pattern analysis. According to the strategy, an algorithm was implemented to rank users with representativeness of the tagging behaviors. Results of the statistical experiments show that the proposed strategy and algorithm can rank tagging users and can be used to discover tagging solutions which are widely accepted by the majority of all the users. Therefore, the proposed strategy and algorithm can be utilized to refine tags for describing items.