Detecting local communities in complex network via the optimization of interaction relationship between node and...

Area of Science:

• Network Science
• Data Mining
• Computational Social Science

Background:

• Local community detection algorithms aim to identify optimal communities around a specific node.
• Existing methods often face challenges with seed selection and accurately measuring node-community relationships.

Purpose of the Study:

• To develop and evaluate a novel local community detection algorithm, OIRLCD.
• To address the seed deviation problem and enhance the measurement of node-community interactions.

Main Methods:

• Introduced an improved seed selection technique.
• Utilized various similarity indices to quantify node-community interaction relationships.
• Developed algorithms based on different similarity indices to optimize these relationships.
• Conducted comparative experiments on real-world and artificial networks.

Main Results:

• The OIRLCD algorithm demonstrates accurate and efficient community detection capabilities.
• Optimization of interaction relationships using node similarity is effective.
• The choice of similarity index significantly impacts the performance of relationship optimization algorithms.

Conclusions:

• The proposed OIRLCD algorithm offers an effective approach to local community detection.
• Optimizing interaction relationships through appropriate similarity indices is crucial for accurate community discovery.
• The algorithm shows strong performance compared to existing methods in various network types.