Grunnleggende konsepter
Existing mean-field methods overestimate the metastable infected fraction in the SIS process due to ignoring correlations, leading to inaccuracies in sparse graphs. New methods incorporating correlations provide more accurate approximations.
Sammendrag
The article discusses the challenges posed by the SIS process on graphs, focusing on the Erdös-Rényi graph. It introduces new methods that consider correlations for more accurate predictions of the infected fraction in metastable states. The study highlights the impact of neighbor correlations and degree distributions on infection rates.
Introduction of SIS process and its significance.
Challenges with existing mean-field methods.
Proposal of new methods considering correlations.
Importance of degree distribution and neighbor correlations.
Simulation results illustrating systematic errors in predictions.