Online social network user behavior analysis — With renren case

Abstract

This paper investigated typical user behaviors in RenRen and used a clustering algorithm that assigns users to groups through a distance measure that is computed based on the values of user feature vector. The user feature vector consists of four attributes and we got six user groups from the clustering process. By analyzing the six different user behavior patterns, we considered some strategies for providers to improve their service quality.