A Collaborative Tag Recommendation Based on User Profile

With the increasing popularity of social tagging, services that assist the user in the task of tagging, such as tag recommenders, are more and more required. As we all known, crucial to the performance of a recommendation system is the accuracy of the user profiles used to represent the interests of the users. We propose a tag recommendation based on user profile which represents user preferences by taking user's ranking pairwise tag preferences. Although the dataset is obtained indirectly, the experiments show that the tag recommendation based on the proposed user profile has outperformed the baseline tag recommendation.