Personalized Photo Recommendation By Leveraging User Modeling On Social Network

Abstract

An online social network is a digital representation of the set of human beings on the Internet. Social network services generate large amount of usage data; for example, Facebook defines detailed user profiles, and provides a platform for sharing information with a vast network of friends, and Flickr offers sophisticated ways for sharing and searching for photos. In this paper, We propose cross-domain user profile modeling that acquires background knowledge from Linked Open Data and measures user interests. We infer the user's preferences by analyzing her Facebook profile, and expand it by linking it to Flickr in order to recommend socially relevant photos.