The framework of a people recommender based on a time series of user preferences

In social media, it is important to build a community that includes people who share similar interests and purposes to foster further interaction and communication. In this study, we present a user profile construction method based on a time series of user preferences to allow the recommendation of appropriate people for the community. This method extracts user preferences as time series data by capturing the user's information browsing behavior in three information spaces: (1) a Web document information space, (2) an augmented reality information space, and (3) an interaction information space with applications for mobile devices. Our proposed method suggests potential members of a clique thus providing opportunities for users of social media to notice the implicit interests associated with other users based on their browsing behavior from the past to the current state, as well as to encourage social communication and relationships.