Comprehensive influence propagation modelling for hybrid social network
RIS ID
111561
Abstract
Springer International Publishing AG 2016.The evolution of influencer marketing relies on a social phenomenon, i.e., influence diffusion. The modelling and analysis of influence propagation in social networks has been extensively investigated by both researchers and practitioners. Nearly all of the works in this field assume influence is driven by a single factor, e.g., friendship affiliation. However, influence spread through many other pathways, such as face-to-face interactions, phone calls, emails, or even through the reviews posted on web-pages. In this paper, we modelled the influence-diffusion space as a hybrid social network, where both direct and indirect influence are considered. Furthermore, a concrete implementation of hybrid social network, i.e., Comprehensive Influence Propagation model is articulated. The proposed model can be applied as an effective approach to tackle the multi-faceted influence diffusion problems in social networks.We also evaluated the proposed model in the influence maximization problem in different scenarios. Experimental results reveal that the proposed model can perform better than those considering a single aspect of influence.
Publication Details
Li, W., Bai, Q. & Zhang, M. (2016). Comprehensive influence propagation modelling for hybrid social network. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 9992 LNAI 597-608.