Community Detection in Complex Networks via Clique Conductance

Area of Science:

• Network Science
• Graph Theory
• Data Mining

Background:

• Complex systems are often analyzed using network science.
• Community structure, or dense local interconnections, is a key feature of networks.
• Existing methods often overlook high-order connection patterns.

Purpose of the Study:

• To develop a novel community-detection method.
• To leverage the mathematical properties of cliques for improved community detection.
• To address limitations of existing methods that focus on low-order connections.

Main Methods:

• Developed a new community-detection algorithm based on cliques (local complete subnetworks).
• Incorporated mathematical properties of cliques to enhance detection accuracy.
• Applied the method to both synthetic and real-world network datasets.

Main Results:

• The method demonstrated high fidelity and sensitivity in detecting known community structures.
• Applied to unknown networks, it revealed insightful organizational patterns.
• Guaranteed near-optimal cluster detection in the bipartition case.

Conclusions:

• The clique-based method offers a robust approach to community detection in complex networks.
• It effectively uncovers network organization by considering high-order patterns.
• Provides valuable insights into the structure of complex systems.