Perturb and Combine to Identify Influential Spreaders in Real-World Networks

TL;DR

This paper introduces a Perturb and Combine (P&C) method for network analysis, which enhances influential spreader detection by creating multiple graph perturbations, applying scoring functions, and aggregating results, leading to improved stability and performance.

Contribution

The paper presents the first P&C procedure for networks, inspired by bagging, that improves influential spreader detection by reducing bias and increasing robustness.

Findings

P&C significantly improves spreader detection accuracy.

Parallelization allows near-zero additional computational cost.

Bias-variance analysis shows P&C mainly reduces bias.

Abstract

Some of the most effective influential spreader detection algorithms are unstable to small perturbations of the network structure. Inspired by bagging in Machine Learning, we propose the first Perturb and Combine (P&C) procedure for networks. It (1) creates many perturbed versions of a given graph, (2) applies a node scoring function separately to each graph, and (3) combines the results. Experiments conducted on real-world networks of various sizes with the k-core, generalized k-core, and PageRank algorithms reveal that P&C brings substantial improvements. Moreover, this performance boost can be obtained at almost no extra cost through parallelization. Finally, a bias-variance analysis suggests that P&C works mainly by reducing bias, and that therefore, it should be capable of improving the performance of all vertex scoring functions, including stable ones.