The reconstruction of complex networks with community structure

Area of Science:

• Network Science
• Computational Biology
• Computer Science

Background:

• Link prediction is crucial for understanding network structures and dynamics.
• Current methods often overlook the importance of community structures in networks.
• Existing approaches struggle to accurately reconstruct networks, especially inter-community links.

Purpose of the Study:

• To evaluate the performance of existing link prediction methods in network reconstruction.
• To address the limitations of current methods in identifying community structures.
• To propose and validate a novel community-based link prediction approach.

Main Methods:

• Applied several representative link prediction methods to network reconstruction tasks.
• Introduced a new community-based link prediction method.
• Evaluated prediction accuracy and effectiveness in reconstructing inter-community links.

Main Results:

• Existing link prediction methods demonstrated poor performance in reproducing network properties.
• These methods failed to identify links connecting different communities.
• The proposed community-based method achieved high prediction accuracy.
• The new method effectively reconstructed inter-community links.

Conclusions:

• Community structure is vital for accurate network reconstruction.
• Existing link prediction techniques are insufficient for capturing community-based network properties.
• The proposed community-based method offers a significant improvement for network reconstruction, particularly for inter-community links.