A Steady State Model for Graph Power Laws

Title: A Steady State Model for Graph Power Laws

Abstract: This research article introduces a new web graph model that generates power law graphs without requiring incremental growth. The model, known as the Steady State (SS) model, operates by repeatedly removing and adding edges to a sparse random graph. The authors compare the SS model with other existing models and provide evidence that it can better capture web graph clustering behavior.

Main Research Question: Can a web graph model generate power law graphs without requiring incremental growth?

Methodology: The researchers developed the Steady State (SS) model, which involves repeatedly removing and adding edges to a sparse random graph. They compared the SS model with other existing models, such as the Waxman model, the GT-ITM tool, and Palmer and Steffan's power law degree generator. They also compared the models with actual web data to determine which one most accurately represents web graph structures.

Results: The researchers found that the SS model can generate power law graphs without requiring incremental growth. They also developed an easily computable graph property that allows them to capture cluster information in a graph without enumerating all possible subgraphs. This property helped them compare the models and provided evidence that the SS model can better capture web graph clustering behavior.

Implications: The Steady State (SS) model provides a new approach to generating power law graphs in web graph models. It does not require incremental growth, which makes it a more realistic model for evolving web graph structures. The model's ability to better capture web graph clustering behavior makes it a valuable tool for understanding the complex structures of the World Wide Web.

Link to Article: https://arxiv.org/abs/0204001v1 Authors: arXiv ID: 0204001v1