Physics > Physics and Society
[Submitted on 9 Dec 2008 (v1), revised 11 Dec 2008 (this version, v2), latest version 11 Mar 2009 (v4)]
Title:Stability of graph communities across time scales
View PDFAbstract: The complexity of networks in biomedicine, the social sciences and engineering makes it desirable to find natural partitions into communities that can act as simplified descriptions and provide insight into the structure and function of the overall system. Although the arsenal of clustering methods is extensive, there is a lack of measures to quantify and rank how good partitions are. We show here that the quality of a partition can be measured in terms of its stability, defined as the magnitude of the clustered autocovariance of a Markov process taking place on the graph. Because the stability has an intrinsic dependence on time scales of the graph, it allows us to compare and rank different clusterings at each time and to establish the time spans over which partitions are optimal. The Markov time acts as an intrinsic resolution parameter that allows us to establish a hierarchy of increasingly coarser clusterings, from the finest partition at time zero to typically bi-partitions at time infinity. Our framework also provides a unifying view of several standard clustering measures: modularity and normalized cut size can be interpreted as one-step time measures whereas Fiedler's spectral clustering emerges in the long time limit. We apply our method to characterize the relevance and persistence of partitions over time for several constructive and real networks, including hierarchical graphs and communities in citation networks. We also obtain reduced descriptions of atom level protein structures over different time scales.
Submission history
From: Jean-Charles Delvenne [view email][v1] Tue, 9 Dec 2008 22:32:02 UTC (369 KB)
[v2] Thu, 11 Dec 2008 17:42:06 UTC (369 KB)
[v3] Mon, 22 Dec 2008 15:49:36 UTC (1,888 KB)
[v4] Wed, 11 Mar 2009 15:26:15 UTC (1,888 KB)
Current browse context:
physics.soc-ph
Change to browse by:
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.