Diffusion/Contagion Processes on Social Networks

Area of Science:

• Social network analysis
• Epidemiology
• Diffusion of innovations

Background:

• Understanding the spread of ideas, practices, or diseases (diffusion or contagion) is crucial for public health interventions.
• Several factors influence diffusion, including adopter characteristics, network structure, and threshold distributions.

Purpose of the Study:

• To model the impact of different seeding strategies, network structures, and threshold distributions on diffusion processes.
• To identify key factors that accelerate and broaden the reach of innovations or behaviors.

Main Methods:

• Modeled seven seeding conditions (opinion leadership, bridging, marginal, random).
• Simulated three network structures (random, small-world, scale-free).
• Analyzed four threshold distributions (normal, uniform, beta 7,14, beta 1,2) with a mean threshold of 33%.

Main Results:

• Seeding with high in-degree centrality or inverse constraint nodes led to faster, wider diffusion.
• Random networks showed faster diffusion than scale-free networks.
• Higher standard deviations in threshold distributions accelerated diffusion and increased prevalence.

Conclusions:

• In-degree centrality and inverse constraint are effective seeding strategies for rapid diffusion.
• Network structure and threshold distribution variance significantly influence diffusion dynamics.
• Findings inform public health strategies for behavior change interventions.