Local Community Detection in Large Networks
Author Information
Author(s): Huang Jianbin, Sun Heli, Liu Yaguang, Song Qinbao, Weninger Tim
Primary Institution: School of Software, Xidian University, Xi'an, China
Hypothesis
Can a new algorithm effectively detect local communities in large-scale networks without global information?
Conclusion
The proposed algorithm efficiently detects multiresolution communities from a source vertex or across the entire network.
Supporting Evidence
- The algorithm can identify overlapping communities in multiresolution by adjusting a resolution parameter.
- Experimental results show that the proposed algorithm is efficient and well-behaved in both real-world and synthetic networks.
- The algorithm was tested on various datasets, including social networks and co-purchase networks, demonstrating its effectiveness.
Takeaway
This study introduces a new way to find groups of connected items in big networks, like social media, without needing to know everything about the network.
Methodology
The study presents a local community detection algorithm based on a similarity-based quality function called tightness.
Limitations
The algorithm's performance may vary based on the order of vertex visits, potentially affecting community detection results.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website