Core of communities in bipartite networks

Area of Science:

• Network science
• Data analysis
• Community detection

Background:

• Bipartite networks represent relationships between two distinct sets of entities.
• Identifying core structures within these networks is crucial for understanding complex systems.
• Existing methods may be sensitive to errors or missing data.

Purpose of the Study:

• To develop a robust method for detecting community cores in bipartite networks.
• To assess the accuracy and precision of the proposed community detection technique.
• To evaluate the method's performance on various real-world and artificial networks.

Main Methods:

• Utilizing information from bipartite networks to create statistically validated projected networks.
• Investigating these projected networks to identify cores of communities.
• Assessing robustness using benchmark, coauthorship, and actor-movie networks.
• Measuring partition accuracy and precision with adjusted Rand and Wallace indices.

Main Results:

• The developed method effectively detects highly informative and robust community cores.
• The community detection process demonstrates high precision.
• Statistical validation confirms the robustness of identified cores against network errors.
• Performance evaluation on diverse networks highlights the method's applicability.

Conclusions:

• The proposed methodology provides a precise and robust approach to community core detection in bipartite networks.
• The technique is resilient to errors and missing data, enhancing its practical utility.
• While precision is high, the overall accuracy of the partition may be context-dependent.