Meta-validation of bipartite network projections

TL;DR

This paper introduces a meta-validation method for bipartite network projections that standardizes significance thresholds across different null models, improving the reliability of validated links in complex systems.

Contribution

It proposes a meta-validation approach to identify model-specific significance thresholds, reducing dependency on null model formulations in bipartite network validation.

Findings

Different null model formulations yield varying validation results.

Meta-validation aligns results across models at consistent link densities.

Application to scientific production data demonstrates the method's effectiveness.

Abstract

Monopartite projections of bipartite networks are useful tools for modeling indirect interactions in complex systems. The standard approach to identify significant links is statistical validation using a suitable null network model, such as the popular configuration model (CM) that constrains node degrees and randomizes everything else. However different CM formulations exist, depending on how the constraints are imposed and for which sets of nodes. Here we systematically investigate the application of these formulations in validating the same network, showing that they lead to different results even when the same significance threshold is used. Instead a much better agreement is obtained for the same density of validated links. We thus propose a meta-validation approach that allows to identify model-specific significance thresholds for which the signal is strongest, and at the same time to obtain results independent of the way in which the null hypothesis is formulated. We illustrate this procedure using data on scientific production of world countries.