Community detection in dynamic networks via adaptive label propagation

Area of Science:

• Network science
• Computational social science
• Data mining

Background:

• Community detection is crucial for understanding complex systems.
• Dynamic networks present challenges for traditional static community detection methods.
• Re-computing community structures after each network change is computationally expensive.

Purpose of the Study:

• To propose a novel algorithm for detecting and monitoring communities in dynamic networks.
• To develop an efficient method that minimizes computational cost during network evolution.
• To provide a scalable and accurate solution for real-world dynamic community detection problems.

Main Methods:

• Introduced an adaptive label propagation algorithm (ALPA).
• ALPA leverages historical community information for efficient updates.
• Employs a local label propagation process to minimize computational impact.

Main Results:

• ALPA successfully identifies and tracks dynamic communities in synthetic and real-world networks.
• Demonstrated high quality and accuracy in community detection compared to existing methods.
• ALPA exhibits low computational cost and operates without requiring parameter tuning.

Conclusions:

• ALPA offers a scalable and efficient solution for dynamic community detection.
• The algorithm's parameter-free nature and low complexity make it suitable for real-world applications.
• ALPA provides a promising approach for monitoring evolving network structures.