Group recommendation system for facebook

Abstract

Online social networking has become a part of our everyday lives, and one of the popular online social network (SN) sites on the Internet is Facebook, where users communicate with their friends, join to groups, create groups, play games, and make friends around the world. Also, the vast number of groups are created for different causes and beliefs. However, overwhelming number of groups in one category causes difficulties for users to select a right group to join. To solve this problem, we introduce group recommendation system (GRS) using combination of hierarchical clustering technique and decision tree. We believe that Facebook SN groups can be identified based on their members’ profiles. Number of experiment results showed that GRS can make 73% accurate recommendation.