Abstract
The proliferation of multi-platform network information has expanded communication channels for users, enabling the integration and dissemination of information across both Social Networking Services (SNS)-type apps and Instant Message (IM)-type apps. With the intensification of convergent communication, some users in the two types of apps show active alternation in spreading information to each other's platforms. This study constructs a heterogeneous-homogeneous asymmetric coupling two-layer network information propagation dynamics model. The propagation threshold $R_0$ and the stability of the model are derived theoretically. Numerical simulations confirm the rationality of the propagation threshold and analyze parameters such as cross-platform circulation and user behavior. Results reveal that promoting platform integration improves communication efficiency. Specifically, dual-platform communication by IM-platform spreaders substantially contributes to the growth of SNS-platform spreaders. Furthermore, higher levels of disinterest among dual-platform spreaders inhibit the growth of spreaders and removers, with IM-type apps demonstrating more pronounced effects.
Introduction
In the contemporary digital landscape, information propagation is no longer confined to single platforms. Users frequently navigate between Social Networking Services (SNS) like X (formerly Twitter) and Instant Message (IM) apps like WhatsApp. This "convergent communication" creates a complex ecosystem where information dynamics differ significantly between public hubs and private relationship chains. This research addresses the structural gap in modeling these interactions by proposing a heterogeneous-homogeneous asymmetric coupling bi-layer network model to understand how information "breaks out" from private circles into the public square.
Methods and Model Construction
The study identifies two distinct network structures that define the bi-layer system:
- SNS-type apps: Characterized as heterogeneous networks where "opinion leaders" or hubs drive radial and explosive propagation.
- IM-type apps: Characterized as homogeneous networks where information flows through "relationship chains" in a gradual, multi-cluster fashion.
The researchers developed a compartmental model (Ignorant-Spreader-Remover) adapted for this bi-layer structure. Unlike symmetric models where every node exists in both layers, this approach utilizes "bridge nodes"—users active on both platforms—to model asymmetric coupling.
The model's dynamics are governed by several key parameters representing transition probabilities between states and across platforms, as detailed in the table below.
The propagation threshold ($R_0$) was derived using the next-generation matrix technique, providing a mathematical tipping point for viral spread. The model was validated using artificial networks (BA scale-free for SNS and WS small-world for IM) and real-world datasets, including Gnutella peer-to-peer and Email networks.
Key Results and Insights
Numerical simulations and real-world data analysis yielded several critical insights into cross-platform dynamics:
- The Bridge Catalyst: Platform integration significantly accelerates information flow. Information in a coupled network reaches its peak density much faster than in isolated, independent networks.
- The IM-to-SNS Pipeline: One of the most striking findings is that IM-platform spreaders contribute more substantially to the growth of SNS-platform spreaders than vice versa. When information moves from a trusted private circle into a public hub, it finds amplifiers that can broadcast it to thousands instantly.
- The "Disinterest" Lever: Increasing the "disinterest rate" ($\lambda$) among dual-platform users is the most effective way to curb rumor growth. This effect is particularly pronounced in the IM layer, which is more sensitive to user turnover and state changes due to its homogeneous structure.
- Threshold Verification: Simulations confirmed that when $R_0 > 1$, information successfully reaches a stable state of awareness across both platforms, whereas $R_0 < 1$ leads to the information dying out.
Conclusion
This study demonstrates that while platform integration improves communication efficiency, it also makes the digital ecosystem more susceptible to the rapid spread of information. The derived propagation threshold $R_0$ serves as a vital tool for predicting and managing information outbreaks. Future research will focus on incorporating user demographics and directional social ties to further refine the model's predictive power in real-world scenarios.
Source: Wang, Y., et al. (2024). Information propagation dynamics on heterogeneous-homogeneous coupling bi-layer networks. Scientific Reports. DOI: 10.1038/s41598-024-80998-8
Summary updated