Network and motif analysis of House of Representatives co-authorship network

Abstract

Co-authorship networks were constructed from the House of Representatives for the 8th to 14th Congress. Network properties and motif-frequency were then analyzed over time. Moreover, properties of unweighted and weighted versions of the co-authorship networks were compared. The three-node motif, which is not fully-connected, were found to occur the most in all networks. For the unweighted and weighted network comparison, weighted networks appear to have larger values of betweenness centrality, degree and shortest path length. However, the clustering coefficient of the unweighted and weighted versions showed no differences. Lastly, motif detection revealed that top subgraphs do not have high degrees or connections.