Abstract:In online social networks such as Facebook, people receive friend recommendations that are based usually on common friends or similar profile such as having the same interest or coming from the same company. However, people receive friend spam in which they do not know why they should add this friend. If we can record the physical context then we can determine how you met that person, and use that for recommending that person to you. In this paper, we create a friend recommendation system using proximity encounters and meetings as physical context called Encounter Meet. We conduct a user study to examine whether physical context-based friend recommendations is better than common friends. Results show that on average, the Encounter Meet algorithm recommended more friends to participants that they added and more recommendations were ranked as good, compared with the common friend algorithm. The results can be used to help design context-aware recommendations in physical environments.

Using Physical Context in a Mobile Social Networking Application for Improving Friend Recommendations