Influence maximization for effective advertisement in social networks: Problem, solution, and evaluation

Abstract

As the number of people using social network services (SNS) increases significantly, companies start to use SNS as a marketing tool. For the reason, an advertisement agent recommendation has been introduced, which selects and recommends advertisement agents who effectively advertize goods of a company in SNS. To address the problem of advertisement agent selection, we propose a multi-state diffusion model. By applying our multi-state diffusion model to existing methods for influence maximization, we could solve the advertisement agent selection problem effectively. In evaluation, we show that the advertisement agents selected by the proposed approach have higher influence spread than the advertisement agents selected by existing methods. In addition, by conducting user study, we confirm that the proposed approach is effective and thus could be used in real-world applications.