Positive influence maximization in signed social networks under independent cascade model
- PDF / 1,203,627 Bytes
- 17 Pages / 595.276 x 790.866 pts Page_size
- 55 Downloads / 288 Views
(0123456789().,-volV)(0123456789().,-volV)
FOUNDATIONS
Positive influence maximization in signed social networks under independent cascade model Jun Sheng1,2 • Ling Chen1 • Yixin Chen3 • Bin Li1 • Wei Liu1
Ó Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract For existing methods for positive influence maximization in signed networks, two factors prevent them from getting highquality results. First, very few researchers consider the critical effect of negative edges on influence propagation. Second, most of those methods use Monte Carlo simulation to estimate the influence propagating of each candidate seed set. Such time-consuming simulation process hinders the application of those methods in solving real-world problems. Motivated by these limitations, this study investigates the problem of positive influence maximization in competitive signed networks. First, an opposite influence propagating model is defined by a set of propagation rules, where negative links play a more critical role than the positive ones. Second, an influence propagation function is defined to estimate the positive influence propagating of a seed set. Using such influence propagation function, the process of simulation can be avoided, and the computation time can be reduced greatly. An algorithm is presented to select the seed nodes which can obtain the largest positive influence spreading in the signed network. The algorithm employs the greedy strategy to sequentially select the seed nodes according to their spreading increments, which are estimated by the influence propagation function. Experimental results on real-world social networks show that our algorithm consistently outperforms the state-of-the-art in terms of solution quality and is several orders of magnitude faster than other methods. Keywords Signed social network Positive influence maximization Independent cascade model Negative influence
1 Introduction
Communicated by A. Di Nola. & Ling Chen [email protected] Jun Sheng [email protected]; [email protected] Yixin Chen [email protected] Bin Li [email protected] Wei Liu [email protected] 1
College of Information Engineering, Yangzhou University, Yangzhou, China
2
School of Information Engineering, Yangzhou Polytechnic College, Yangzhou, China
3
Department of Computer Science, Washington University in St. Louis, St. Louis, USA
Recently, with the booming development of social networks, such as Twitter, Google? and China WeiBo, network analyses have become a hotspot of research in many areas, such as information diffusion (Fu et al. 2019; Gao et al. 2019; Hosseini-Pozveh et al. 2019), community detection (Ahmad et al. 2019), direction recovery (Wen et al. 2019), metric learning (Li et al. 2018a, b), network structure analyzing (Yu et al. 2019) and link prediction (Ahmed et al. 2018). Social network created a great impact on the information propagation and has become an ideal platform for people to exchange opinions and to propagate information. For example, a merchant of Amazon may promote a new product t
Data Loading...
