SEIM: Search economics for influence maximization in online social networks

• Published in: Future generation computer systems Vol. 93; pp. 1055 - 1064
• Main Authors: Tsai, Chun-Wei, Liu, Shih-Jui
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.04.2019
• Subjects: Influence maximization problem; Metaheuristic algorithm; Online social networks; Search economics; Influence maximization problem; Search economics; Metaheuristic algorithm; Online social networks
• ISSN: 0167-739X; 1872-7115
• DOI: 10.1016/j.future.2018.08.033

The influence of online social networks (OSN), which can be regarded as part of our life, is evident today. As expected, a great deal of useful information about the humans is hidden in the data, such as interpersonal relationship and personal preference. The influence maximization problem (IMP) is one of the well-known problems in this research domain that has attracted the attention of researchers from different disciplines in recent years. One of the reasons is that it can speed up the propagation of information in OSN if we can find out users that have maximum influence on other users. However, traditional rule-based and heuristic algorithms may not be able to find useful information out of these data because the data are generally large and complex. Although many recent studies attempted to use metaheuristic algorithms to solve the IMP, there is still plenty of room for improvement. The proposed algorithm, called search economics for influence maximization (SEIM), is motivated by the concept of return on investment to design its search strategies. As far as the proposed algorithm is concerned, the search strategy of SEIM is like making a good plan to determine how to invest the high potential investment subjects (i.e., regions) in the market (i.e., search space). The experimental results show that the proposed algorithm is significantly better than the other state-of-the-art influence maximization problem algorithms compared in this paper in terms of the quality of the end result and the number of objective function evaluations.