Social contagions with communication channels alternation on multiplex networks

TL;DR

This paper investigates how communication channel alternation (CCA) affects social contagion dynamics on multiplex networks, revealing that CCA delays but suppresses spreading and causes hybrid growth patterns in adoption size.

Contribution

It introduces a non-Markovian model incorporating CCA and develops a generalized edge-based method to analyze spreading dynamics on multiplex networks.

Findings

CCA delays behavior spreading but does not change final adoption size

CCA suppresses overall behavior spreading

Hybrid growth patterns observed in coupled networks with CCA

Abstract

Internet communication channels, e.g., Facebook, Twitter, and email, are multiplex networks that facilitate interaction and information-sharing among individuals. During brief time periods users often use a single communication channel, but then communication channel alteration (CCA) occurs. This means that we must refine our understanding of the dynamics of social contagions. We propose a non-Markovian behavior spreading model in multiplex networks that takes into account the CCA mechanism, and we develop a generalized edge-based compartmental method to describe the spreading dynamics. Through extensive numerical simulations and theoretical analyses we find that the time delays induced by CCA slow the behavior spreading but do not affect the final adoption size. We also find that the CCA suppresses behavior spreading. On two coupled random regular networks, the adoption size exhibits hybrid growth, i.e., it grows first continuously and then discontinuously with the information transmission probability. CCA in ER-SF multiplex networks in which two subnetworks are Erd\H{o}s-R\'{e}nyi (ER) and scale-free (SF) introduces a crossover from continuous to hybrid growth in adoption size versus information transmission probability. Our results extend our understanding of the role of CCA in spreading dynamics, and may elicit further research.