This website contains resources and materials to support our work on faster computation of percolations in complex networks.
We provide source code of several k-clique percolation implementations used in our work, in C++ and Python.
Algorithm 1:
A python implementation is available.
Usage: python CliquePercAlgorithm1.py inputFileName k outputFileName
where 'inputFileName' is a graph edgelist, tab separated, one edge per line, undirected, unweighted, with no self loops.
k is the value of k required.
The output will be written to outputFileName, as a space separated list of communities, one community per newline.
The python code may easily be modified to accept or output different formats, and uses the networkx library.
A C++ implementation of algorithm 1 is also available.
Algorithm 2:
A C++ implementation is available.
Datasets:
The ca-CondMat graph, benchmarked in the paper, is available from the SNAP network project here:
http://snap.stanford.edu/data/ca-CondMat.html
The GN graphs were generated using software provided here:
https://sites.google.com/site/santofortunato/inthepress2
Specifically:
https://sites.google.com/site/santofortunato/binary_networks.tar.gz?attredirects=0
Other algorithms:
The authors of CFinder have made a binary (no source code) CFinder implementation available here:
The authors of SCP have made an unoptimised python implementation available here:
http://www.lce.hut.fi/~mtkivela/kclique.html
We use a binary implementation provided by the SCP authors for our benchmarks.
A binary SCP implementation is also available on that page.
Acknowledgements:
We are grateful to the authors of all of the above resources, for providing them on-line.
The authors would like to thank our colleague Conrad Lee, for valuable discussions and advice.
This work is supported by Science Foundation Ireland under grant 08/SRC/I1407: Clique: Graph and Network Analysis Cluster.
Misc:
We are happy to answer questions on this data or software, please e-mail us with queries.