Node-motifs: profiling nodes through motif-based analysis in social networks

Abstract In this paper we propose a general framework and algorithms for assessing the role and importance of nodes within a complex network. We introduce a node classification approach that allows a node’s role to be characterised through its pos- sible participation in any 3 or 4-node connected network motif. This classification approach gives insight into how nodes contribute to local flooding, and therefore also gives insight into the role of nodes in complex contagion. We define those nodes that are over represented relative to some comparable class of random graph as node-motifs. This gives a new approach for intra-network comparison of nodes and understanding of their relative importance to the network, with potentially widespread applications.
  • Roger Whitaker (Cardiff)
  • Liam Turner (Cardiff)
  • Lauren Hudson (Cardiff)
  • Jian Li (UMass)
  • Don Towsley (UMass)
  • Diane Felmlee (PSU)
Date Sep-2018
Venue 2nd Annual Fall Meeting of the DAIS ITA, 2018