AdityaThaokar / Community-Detection-Empirical-Study

We have compared the different community detection algorithms for multiple real-world social networks with some standard datasets.

Community Detection Empirical Study

Datasets :

Karate Club : http://www-personal.umich.edu/~mejn/netdata/karate.zip

Dolphin : http://www-personal.umich.edu/~mejn/netdata/dolphins.zip

Jazz Musicians : http://deim.urv.cat/~alexandre.arenas/data/xarxes/jazz.zip

File q1.py

It contains the following statistics describing the datasets :

- number of nodes n,

- number of edges m,

- average path lengths d ̄

- average clustering coefficient C

File q2-*.py

It computes the following for each dataset :

- betweenness-based clustering using the Girvan-Newman algorithm,

- modularity-based clustering (modularity maximization),

- spectral clustering (using the graph Laplacian).

File q3-*.py

It computes the following for each dataset :

- number of clusters found

- modularity score for this clustering

- run time of the algorithm

Contributors :

Thaokar Aditya Mukund - 2020H1030132P

Kaushal Rajan Hatwar - 2020H1030122P