A tree-network model for mining short message services seed users and its empirical analysis

Identifying short message services (SMSs) seed users helps to discover the information’s originals and transmission paths. A tree-network model was proposed to depict the characteristics of SMS seed users who have such three features as “ahead of time”, “mass texting” and “numerous retransmissions”. For acquiring the established network model’s width and depth, a clustering algorithm based on density was adopted and a recursion algorithm was designed to solve such problems. An objective, comprehensive and scale-free evaluation function was further presented to rank the potential seed users by using the width and the depth obtained above. Furthermore, the model’s empirical analysis was made based on part of the Shenzhen’s cell phone SMS data in February of 2012. The model is effective and applicable as a powerful tool to solve the SMS seed users’ mining problem.

► This paper grasps three features of seed users and designs a tree-network.
► We propose an evaluation model to rank the users based on the tree-network model.
► This paper also makes an empirical analysis to apply the new method.
► The evaluation function presented is objective, scale-free and of good discrimination.
► The model proposed is proper and functional for mining SMS seeders.