Communities

Required readings

• G. Palla et al., Uncvoering the overlapping community structure of complex networks in nature and society, Nature, 2005.
• Y.-Y. Ahn et al., Link communities reveal multiscale complexity in networks, Nature, 2010.
• B. Ball et al., Efficient and principled method for detecting communities in networks, PRE, 2011.

Further readings

• T.S. Evans and R. Lambiotte, Line graphs, link partitions, and overlapping communities, PRE, 2009 - Independently developed line-graph based approach.
• T.S. Evans, Clique graphs and overlapping communities, Journal of Statistical Mechanics, 2010 - Higher order generalization
• I. Derény et al., Clique percolation in random networks, PRL, 2005 - Analytical study on the properties of clique percolation in random networks.
• J.M. Kumpula et al., A sequential algorithm for fast clique percolation, PRE, 2008 - An improved algorithm of clique percolation. Also provide natural generalization to weighted networks.
• Python implemention of the SCP algorithm - Python source code for the sequential algorithm.
• S. Lehmann et al., Bi-clique communities PRE, 2008 - Extension of clique percolation to bi-partite graphs
• C. Lee et al., Detecting highly overlapping community structure by greedy clique expansion, arXiv, 2010 - Another approach based on cliques.

Questions

1. Why should we care about ‘overlapping communities’?
2. What is a ‘clique’ and what is a ‘k-clique’?
3. What is the definition of a community in the Clique Percolation Method? Why do they define it this way?
4. What’s the drawbacks of CPM?
5. Why community overlap and hierarchical structure are incongruent in node-community paradigm?
6. What are the benefits of ‘link community’ concept? Any drawbacks?
7. What is ‘line graph transformation’? What are the problems of simple line graph transformation? How does link community method improve?
8. What is the main idea of ‘generative modeling framework’? How is it related to the maximum-likelihood framework?