- Scalable Discovery of Best Clusters on Large Graphs, VLDB 2010
- Weighted: yes
- Directed: yes
- Scalable: yes
- Overlap: yes

This algorithm aims to find the 'top clusters' not finding all clusters in the graph. It uses Locality-sensitive hashing. The quality measure is defined as "the average link weight between all nodes in the cluster (including links with weight 0) multiplied by the square root of the cluster size (to allow for a bias towards larger cluster sizes)."

$$ Q_c = \sqrt{|c|} \frac{\sum_{i,j \in c} w_{ij}}{|c| (|c|-1)} $$

# Incoming Links #

## Related Articles (Article 0) #

# Suggested Pages #

- 0.422 Coarse graining
- 0.188 Community structure
- 0.055 Dynamic community structure
- 0.045 Community detectability
- 0.040 Block model
- 0.036 Modular structure of brain
- 0.032 Community evolution
- 0.027 Temporal community structure
- 0.026 Animal social network
- 0.021 Community
- More suggestions...