An Incremental Clustering Approach To Personalized Tag Recommendations

Volumes of user-generated contents have caused the problem of information overload and hindered Internet users from browsing and retrieving information. Social tagging that allows users to annotate resources with free preferred keywords to ease the access to their collecting resources. Though social tagging benefits users managing their resources, it always suffers the problems such as diverse and/or unchecked vocabulary and unwillingness to tag because tags are freely and voluntarily assigned by users. Tag recommender systems, which follow some criteria to select from the tag space the most relevant tags to the user's annotating resource, drastically transfer the tagging process from generation to recognition to reduce user's cognitive effort and time. This study takes personalized tag recommendation as an incremental clustering problem and proposes a Progressive Expansion-based Tag (PET) recommendation technique. The incremental clustering assumes each object appears in sequence and then is incrementally clustered into either an appropriate existing category or a created new category. The PET technique can classify each resource into multiple categories (i.e., tags) or label it as new. While a resource is labelled as new, it will recommend a set of tags that have been used by other users and are relevant to the target user's practices. Finally, our empirical evaluation results suggest that the proposed PET technique outperforms the traditional popularity-based tag recommendation methods, while the performance rates achieved by both techniques are not satisfying.