Computer Science > Social and Information Networks
[Submitted on 28 Feb 2020 (v1), revised 14 May 2020 (this version, v2), latest version 7 Jan 2021 (v3)]
Title:Modularity-Impact: a Signed Group Centrality Measure for Complex Networks
View PDFAbstract:Group-centrality is an emerging area in Network Science which seeks to quantify the importance of nodes within and between communities. Previous group-centrality measures either have strong theoretical underpinnings in community detection, or scale well, but not both. We propose an extremely scalable measure, modularity-impact, that is directly derived from Newman's modularity. Unlike other measures, modularity-impact is signed, providing additional information about each node's role: hub within its community, or bridge between communities. We analyze our results in a network robustness framework, as it is a generalized version of immunization under SIR models, which prior works have used for evaluation. We show that our measure out-performs previous measures for immunizing strategies in generated networks with strong community structure. Additionally, we show that our measure gives the most effective attack strategy for a million-node infrastructure network, fragmenting it while removing less than 2% of nodes. This is over 8 times more fragile than can be seen with previous group-centralities. We further find that the social media communication network demonstrates extreme robustness to group-centrality attacks, suggesting that user-based interventions to mitigate misinformation diffusion will be ineffective. Finally, we show that our group-centrality measure can be used as a greedy solution to the community deception problem.
Submission history
From: Thomas Magelinski [view email][v1] Fri, 28 Feb 2020 20:35:58 UTC (4,267 KB)
[v2] Thu, 14 May 2020 17:33:26 UTC (2,161 KB)
[v3] Thu, 7 Jan 2021 18:56:59 UTC (2,009 KB)
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.