A multi-objective Algorithm based on Structural Information to Identify Influential Users in Social Networks

Social influence includes the change in behavior or action of individuals in a particular way to meet the demands of a social environment. Research on social influences indicates that accepting or rejecting a new idea by a person depends on the acceptance or rejection of his friends. Therefore, many companies are focused on “word-of-mouth” marketing to take advantage of network effects in advertisement. In such kind of advertisement, a small set of influential users in a social network are selected, such that the highest expansion in the network can be reached with the lowest cost. The set of influential users must be chosen in such a way that they do not activate similar users. In this paper, an optimization model is presented to identify the most valuable users who have little similarity to each other. This model considers two objectives of maximizing profit and minimizing user similarity, simultaneously. Evaluation on real datasets confirms that the proposed method selects a more appropriate set of target users to achieve greater profitability and higher influence power than other methods.