Greedy algorithm for edge-based nested community detection

We propose an edge-based community detection algorithm that finds nested communities of a given graph. The communities are defined as the subgraphs induced by the edges of the same label and these edges together fulfill the property of network nestedness. Our method compares only possibly nested pairs of nodes and assigns all their edges to either common or different communities, realizing nested subgraphs. Finally, the algorithm removes superfluous communities in a post-processing step. We inspect the algorithm’s performance on a set of host-parasite networks and show the correlation between mean community size and the discrepancy nestedness measure. Since the algorithm’s performance is adjustable through a threshold parameter, we also investigate the effects of the parameter on the number of iterations and the obtained community structure.