Tagpref: User Preference Modeling by Social Tagging

User preference modeling is the basis for the provision of personal information services because it can reflect the user's characteristics and interests. Tagging is a popular online activity that allows a user to discover, describe, and organize Internet contents. Tagging provides a new means of creating user preference. Tags explain and describe items so that a user's attitude toward a tag is his or her attitude toward an aspect of the item. In this paper we design Tagpref inspired by the method of establishing user preference with tags. The user's preference was described by the user's preference degrees for different tags. Different algorithms were proposed and evaluated for taggers and non-taggers on Movie Lens based on direct and indirect behavior. Evaluation results show that the proposed algorithms can accurately predict a user's tag preference. The user's attitude toward tag preference was also analyzed. Users' true preference was determined by an online user survey.