Improving the accuracy of Tagging Recommender System by Using Classification

Collaborative tagging system has become more and more popular and recently achieved widespread success due to flexibility and conceptual comprehensibility of tagging systems. Recommender system has the access to adopt tagging systems to achieve better performance. In this paper we consider that the items can be categorized into different classifications in which users show different interests. Here we adopt a two-step recommender method called TRSUC (Tagging Recommender Systems by Using Classification) which can be described as Inner-Class Recommender or Global Recommender in which we use tag as the intermediary entity between user and item. The experiment using MovieLens as dataset shows that we acquire better results than the recommender algorithms without classifying the items.