Reproduce “Preferential Attachment in Online Networks: Measurement and Explanations” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Jérôme Kunegis, Marcel Blattner, Christine Moser
An extensive empirical study of 47 online networks reveals that preferential attachment is predominantly nonlinear, with networks consistently exhibiting either sublinear or superlinear growth patterns depending on their category, challenging the long-held assumption of strict linearity. The observed preferential attachment exponent serves as a robust metric to differentiate network types and their underlying formation mechanisms.
Key insights
- Preferential attachment in online networks is predominantly nonlinear (β ≠ 1), directly challenging the conventional assumption of a strictly linear relationship (β = 1).
- Networks consistently cluster into two distinct categories based on their β exponent: sublinear (β < 1), primarily seen in rating, communication, and folksonomy networks, and superlinear (β > 1), common in explicit/implicit interaction and large Wikipedia edit networks.
- The preferential attachment exponent (β) serves as a novel and reliable numerical network measure capable of characterizing and differentiating various types of network growth dynamics, offering insights beyond static degree distributions.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.