Generating Bipartite Networks with a Prescribed Joint Degree Distribution

Area of Science:

• Network Science
• Graph Theory
• Computational Social Science

Background:

• Bipartite networks model interactions between two distinct sets of entities.
• Real-world networks often exhibit correlations between node degrees and their neighbors' degrees.
• Preserving these structural properties is crucial for accurate network modeling.

Purpose of the Study:

• To introduce novel algorithms for constructing bipartite networks.
• To ensure the generated networks preserve both degree and joint-degree distributions.
• To evaluate the effectiveness of these algorithms in capturing network structure.

Main Methods:

• Development of bipartite 2K (B2K) algorithms.
• Generation of an ensemble of networks with prescribed degree sequences.
• Preservation of the joint-degree distribution (degree of neighbors for nodes of same degree).
• Utilizing the NetworkX software environment for analysis.

Main Results:

• B2K algorithms successfully preserve specified degree and joint-degree distributions.
• Comparison with other methods on a romance network dataset.
• Observed preservation of additional network properties (e.g., clustering coefficients) in low-degree networks.

Conclusions:

• B2K algorithms provide a robust method for constructing realistic bipartite networks.
• The joint-degree distribution is a critical property to preserve, especially in sparse networks.
• These algorithms enhance the fidelity of network models for systems with low average node degrees.