Information Propagation Model in Hybrid Networks

TL;DR

This paper introduces a hybrid network model combining scale-free and small-world features with various epidemic models to better simulate real-world information propagation, validated through theoretical analysis and simulations.

Contribution

A novel hybrid network model integrating multiple epidemic processes and features like blockbuster effect to closely mimic real network propagation dynamics.

Findings

Network degree distribution follows a power law.

Model closely matches real network propagation patterns.

Hybrid model effectively captures explosive spread phenomena.

Abstract

It is in practice impossible to describe the topology of a real network or its message propagation process using a single dynamic model. To address this issue, we constructed a new hybrid network model based on scale-free (SF), small-world (SW) features that functions as closely as possible to a real network. And the hybrid propagation model is constructed with susceptible-infected-susceptible (SIS), susceptible-infected-recovered (SIR) and susceptible-infected-recovered-susceptible (SIRS) model mixed in arbitrary proportions. The model applies the concepts of blockbuster effect and implicit node edges to reflect explosive spread as a significant characteristic of information propagation. A theoretical analysis and derivation of the new model in which hybrid networks were simulated revealed that the network degree distribution closely follows a power law. Using an improved similarity function to define the degree of closeness to real network cases, the proposed model was shown to be valid and very close to a real network.