Identifying Rumor Sources Using Dominant Eigenvalue of Nonbacktracking Matrix

TL;DR

This paper introduces a novel eigenvalue-based algorithm for identifying rumor sources in networks, outperforming existing methods in accuracy by leveraging the dominant eigenvalue of a reduced nonbacktracking matrix.

Contribution

The paper proposes a new eigenvalue-based approach for rumor source detection and a reduced-complexity approximation, improving accuracy over existing algorithms.

Findings

Higher accuracy in source identification on synthetic networks

Effective performance on real-world networks

Reduced computational complexity with the approximation

Abstract

We consider the problem of identifying rumor sources in a network, in which rumor spreading obeys a time-slotted susceptible-infected model. Unlike existing approaches, our proposed algorithm identifies as sources those nodes, which when set as sources, result in the smallest dominant eigenvalue of the corresponding reduced nonbacktracking matrix deduced from message passing equations. We also propose a reduced-complexity algorithm derived from the previous algorithm through a perturbation approximation. Numerical experiments on synthesized and real-world networks suggest that these proposed algorithms generally have higher accuracy compared with representative existing algorithms.