On the accuracy of message-passing approaches to percolation in complex networks

Abstract

The Message-Passing Approach (MPA) is the state-of-the-art technique to obtain quasi-analytical predictions for percolation on real complex networks. Besides being intuitive and straightforward, it has the advantage of being mathematically principled: it is exact on trees, while yielding generally good predictions on networks containing cycles as do most real complex networks. Here we show that the MPA does not perform its calculations on some ill-defined tree-like approximation of the network, as its formulation leads to believe, but rather considers a random network ensemble in which the original network is cloned and shuffled an infinite number of times. We conclude that the fact that the MPA is exact on trees does not imply that it is nearly exact on tree-like networks. In fact we find that the closer a non-tree network is to a tree, the worse the MPA accuracy becomes.

7 pages, 4 figures

References in corpus (10)

• A message passing approach for general epidemic models
• Percolation on sparse networks
• Percolation in real interdependent networks
• Scalable detection of statistically significant communities and hierarchies, using message-passing for modularity
• Dynamic message-passing equations for models with unidirectional dynamics
• Mapping the Structure of Directed Networks: Beyond the "Bow-tie" Diagram
• Spectra of random networks with arbitrary degrees
• Network cloning unfolds the effect of clustering on dynamical processes
• Percolation in real multiplex networks
• Unsupervised feature learning from finite data by message passing: discontinuous versus continuous phase transition