Modularity and community detection in bipartite networks

Area of Science:

• Network Science
• Graph Theory
• Computational Biology

Background:

• Network modularity quantifies community structure relative to a null model.
• Bipartite networks, common in biology and social sciences, require specialized analysis.
• Existing modularity measures may not adequately capture the structure of bipartite systems.

Purpose of the Study:

• To define a null model and modularity metric specifically for bipartite networks.
• To develop and present an algorithm for identifying modules in bipartite networks.
• To validate the algorithm's effectiveness on real-world network data.

Main Methods:

• Definition of a null model appropriate for bipartite networks.
• Introduction of a bipartite modularity measure using a modularity matrix B.
• Development of a module-detection algorithm leveraging the eigenspectrum of B and mutual induction between network parts.

Main Results:

• Key properties of the eigenspectrum of the bipartite modularity matrix B were identified.
• An algorithm for bipartite network module detection was successfully developed.
• The algorithm demonstrated efficacy in identifying modular structures in real-world bipartite network datasets.

Conclusions:

• The proposed bipartite modularity provides a robust framework for analyzing community structure in bipartite networks.
• The developed algorithm effectively identifies modules by considering the interdependence of network partitions.
• This work offers a valuable tool for understanding the organization of complex bipartite systems.