Network properties and motif characterization of evolved social networks

Abstract

Different real social networks such as friendship and co-authorship networks were characterized over time and its motifs were tabulated as the networks evolved. The fully-connected 3-node sub-graph was found to significantly appear the most in all networks. We have seen that as the motif become more connected, the motif-frequency for the influence and friendship networks decrease while the motif-frequency for the Senate co-authorship networks increase. The significantly abundant network motifs were also indentified using randomized version of the networks and some parameter cutoffs.