Discovering Patterns in Bipartite Networks

Bipartite networks constitute an especially important type of network that can be widely applied to model and study several intricate systems in many scientific areas, being also theoretically related to several network models and concepts. However, given a bipartite network the identification of specific recurrent patterns of interest often represents a substantial challenge. In the present work, we apply the coincidence methodology for translating data into networks as a means for automatically identifying repeating interconnection patterns in given bipartite networks. The important issue of normalization of the links strength is also addressed. Though the method is illustrated with respect to foodweb networks, its application is general and encompasses many other subjects and areas.