Parts | Days | Selection | Search | Updates | Downloads | Help
DY: Fachverband Dynamik und Statistische Physik
DY 33: Networks, From Topology to Dynamics - Part II (joint session SOE/ DY/ BP)
DY 33.3: Talk
Thursday, April 3, 2014, 12:45–13:00, GÖR 226
Analysis of local network structure by node-specific triadic Z-score profiles — •Marco Winkler and Jörg Reichardt — Institute for Theoretical Physics, University of Würzburg, Germany
Over the last decade so called network motifs have attracted high attention. A motif is a subgraph pattern that appears significantly more often than in a random network with the same degree distribution as the original one. Triadic Z-score profiles, Z→, assign every possible triadic subgraph pattern i a score Zi, corresponding to the magnitude of over-/underrepresentation of the pattern compared to the random null model. These Z-score profiles are a common tool to analyze complex networks.
However, triad patterns are not necessarily homogeneously distributed over the network. Therefore, we introduce the concept of node-specific Z-scores. For the node-specific Z-score profile, Z→α, of a node α, only the triads it participates in are taken into account. The node-specific Z-score profiles can then be used for classification of a network’s vertices into different structural groups. We present results for various real-world data sets including neural networks and transcription networks.