2010/02/09 by Conrad Lee, Lee, Conrad, Fergal Reid +5 · 1 citation
Biochemistry, Genetics and Molecular Biology · Computer Science · Physics and Astronomy · #Advanced Graph Neural Networks #Bioinformatics and Genomic Networks #Complex Network Analysis Techniques #Data Analysis #FOS: Physical sciences #H.2.8 #Physics and Society (physics.soc-ph) #Statistics and Probability (physics.data-an)
paper · pdf · doi:10.48550/arxiv.1002.1827
openalex publication_date 2010/02/09 · openalex created_date 2022/10/02 · openalex updated_date 2026/07/28
In complex networks it is common for each node to belong to several\ncommunities, implying a highly overlapping community structure. Recent advances\nin benchmarking indicate that existing community assignment algorithms that are\ncapable of detecting overlapping communities perform well only when the extent\nof community overlap is kept to modest levels. To overcome this limitation, we\nintroduce a new community assignment algorithm called Greedy Clique Expansion\n(GCE). The algorithm identifies distinct cliques as seeds and expands these\nseeds by greedily optimizing a local fitness function. We perform extensive\nbenchmarks on synthetic data to demonstrate that GCE's good performance is\nrobust across diverse graph topologies. Significantly, GCE is the only\nalgorithm to perform well on these synthetic graphs, in which every node\nbelongs to multiple communities. Furthermore, when put to the task of\nidentifying functional modules in protein interaction data, and college dorm\nassignments in Facebook friendship data, we find that GCE performs\ncompetitively.\n