Comparing Community-aware Centrality Measures in Online Social Networks

TL;DR

This paper systematically compares various community-aware centrality measures in online social networks, revealing which are most effective for identifying influential nodes during epidemic spreading.

Contribution

It provides the first comprehensive evaluation of community-aware centrality measures using real-world data and the SIR model, highlighting the most accurate methods.

Findings

K-shell with Community and Community-based Centrality are most accurate.

Epidemic transmission rate has little impact on measure effectiveness.

Community-aware measures outperform traditional centrality metrics.

Abstract

Identifying key nodes is crucial for accelerating or impeding dynamic spreading in a network. Community-aware centrality measures tackle this problem by exploiting the community structure of a network. Although there is a growing trend to design new community-aware centrality measures, there is no systematic investigation of the proposed measures' effectiveness. This study performs an extensive comparative evaluation of prominent community-aware centrality measures using the Susceptible-Infected-Recovered (SIR) model on real-world online social networks. Overall, results show that K-shell with Community and Community-based Centrality measures are the most accurate in identifying influential nodes under a single-spreader problem. Additionally, the epidemic transmission rate doesn't significantly affect the behavior of the community-aware centrality measures.