Community-Based Acceptance Probability Maximization for Target Users on Social Networks

• Published in: Algorithmic Aspects in Information and Management Vol. 11343; pp. 293 - 305
• Main Authors: Yan, Ruidong, Zhu, Yuqing, Li, Deying, Wang, Yongcai
• Format: Book Chapter
• Language: English
• Published: Switzerland Springer International Publishing AG 2018; Springer International Publishing
• Series: Lecture Notes in Computer Science
• Subjects: Approximate algorithm; Community structure; Seed selection; Social influence; Submodularity
• ISBN: 3030046176; 9783030046170
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-030-04618-7_24

Social influence problems, such as Influence Maximization (IM), have been widely studied. But a key challenge remains: How does a company select a small size seed set such that the acceptance probability of target users is maximized? In this paper, we first propose the Acceptance Probability Maximization (APM) problem, i.e., selecting a small size seed set S such that the acceptance probability of target users T is maximized. Then we use classical Independent Cascade (IC) model as basic information diffusion model. Based on this model, we prove that APM is NP-hard and the objective function is monotone non-decreasing as well as submodular. Considering community structure of social networks, we transform APM to Maximum Weight Hitting Set (MWHS) problem. Next, we develop a pipage rounding algorithm whose approximation ratio is ( $$1-1/e$$ ). Finally, we evaluate our algorithms by simulations on real-life social networks. Experimental results validate the performance of the proposed algorithm.