A Probabilistic Profile Reliability Approach to Improve the Richness of User's Interests Based on Social Information

Abstract

The user requires personalized information that depends on his/her current profile information (needs, interests, preferences, etc.). Interests are the most important information of the user’s profile. There are many sources which may provide beneficial information to model the user’s interests, such as social neighbors’ profiles.However, the latter may provide conflicting interests (erroneous, duplicated, out of date, ambiguous,etc.) and consequently they can be considered as non-reliable sources. In this paper, we propose aprobabilistic approach to handle conflicts by detecting reliable profiles so as to improve the richness of user’sinterests. Our approach is different from those that are previously proposed as it takes into account the organizational aspects of interests in terms of their evolutionary aspect (freshness and popularity) as wellas their semantic relationships. Our experiment was conducted in a social-learning context in order to take the case of the improvement of the learner’s profile content based on his/her social network. The experiment led to satisfactory results.