Influence maximization in multilayer networks based on adaptive coupling degree

TL;DR

This paper introduces a new influence maximization approach in multilayer networks using an adaptive coupling degree heuristic, improving influence spread and efficiency over existing methods.

Contribution

It proposes a novel multilayer independent cascade model and an adaptive coupling degree heuristic for seed node selection, addressing the gap in multilayer influence maximization research.

Findings

The proposed ACD algorithm outperforms baselines in influence spread.

ACD reduces time cost compared to existing methods.

Experimental results on synthetic and real-world networks validate effectiveness.

Abstract

Influence Maximization(IM) aims to identify highly influential nodes to maximize influence spread in a network. Previous research on the IM problem has mainly concentrated on single-layer networks, disregarding the comprehension of the coupling structure that is inherent in multilayer networks. To solve the IM problem in multilayer networks, we first propose an independent cascade model (MIC) in a multilayer network where propagation occurs simultaneously across different layers. Consequently, a heuristic algorithm, i.e., Adaptive Coupling Degree (ACD), which selects seed nodes with high spread influence and a low degree of overlap of influence, is proposed to identify seed nodes for IM in a multilayer network. By conducting experiments based on MIC, we have demonstrated that our proposed method is superior to the baselines in terms of influence spread and time cost in 6 synthetic and 4 real-world multilayer networks.