Finding Statistically Significant Communities in Networks
Author Information
Author(s): Andrea Lancichinetti, Filippo Radicchi, José J. Ramasco, Santo Fortunato
Primary Institution: Institute for Scientific Interchange (ISI), Torino, Italy
Hypothesis
Can we develop a method to detect clusters in networks that accounts for various features like edge directions and overlapping communities?
Conclusion
The OSLOM method effectively detects statistically significant communities in networks, outperforming existing algorithms in various scenarios.
Supporting Evidence
- OSLOM can handle directed graphs and overlapping communities.
- It outperforms existing algorithms on artificial benchmark graphs.
- OSLOM identifies the absence of community structure effectively.
- The method is versatile and applicable to various types of networks.
Takeaway
This study introduces a new method called OSLOM that helps find groups in networks, even when the connections are complicated or shared between groups.
Methodology
The OSLOM method uses local optimization of a fitness function to assess the statistical significance of clusters in networks.
Limitations
The method may find 'homeless' vertices that do not belong to any cluster due to random fluctuations.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website