A Particle Swarm Optimization Based Approach for Finding Similar Users on Facebook

With ever increasing size and use of social networks, there is a large amount of information which users implicitly or explicitly leave behind on social media. This information can be used to identify their personal traits and preferences. In this work, we have proposed a Particle Swarm Optimization (PSO) based approach for clustering users on Facebook (FB) data to identify similar users. Proposed method can be used to recommend people having similar interests. Our results indicate a lesser quantization error than k-means when used for 7 or less clusters. Possible applications of this approach include rope-in exercises for new hires, assigning a resource person to work team, etc.