This program is an implementation of the community detection algorithm described in the paper titled

High quality, scalable and parallel community detection for large real graphs. Arnau Prat-Pérez, David Dominguez-Sal, Josep-Lluis Larriba-Pey - WWW 2014.

SCD uses CMake 2.8.2 or greater to compile. In order to build SCD, move to SCD directory and type:

cd build
cmake -DCMAKE_BUILD_TYPE=Release ..
make

This will create a build directory into the SCD folder tree, and configure and build SCD in Release mode. In order to compile SCD in Debug mode, please replace the last two lines of the snippet above by:

cmake -DCMAKE_BUILD_TYPE=Debug ..
make

To execute SCD type, move to the build folder and type:

./scd -f [network file name]

where the [network file name] contains the network with an edge per line, and each edge is represented as a pair of numeric identifiers. IMPORTANT: each edge is interpreted as an undirected edge and can only appear once. For example, if "1 2" appears in the file, then "2 1" cannot appear too. The next snipped shows a valid network file:

1 2
4 5
3 4
1 3

As an example, type:

./scd -f ./network.dat

This will run the program, and will output the communities found at network.dat to "./communities.dat", which contains a community per line, represented as a list of identifiers. network.dat contains a network composed by two cliques of size 5 linked by a single edge. Therefore, communities.dat outputs:

1 2 3 4 5
6 7 8 9 10

Now, we summarize the options of the program:

• -f [netork file name] : Specifies the network file.
• -o [output file name] : Specifies the output file name (DEFAULT="communities.dat").
• -n [number of threads]: Specifies the number of threads to run the algorithm (DEFAULT=maximum available cores).
• -l [lookahead]: Specifies the number of lookahead iterations to look before terminating the optimization process (DEFAULT=5).
• -p [partition file name]: Specifies the initial partition to refine from (Optional).

Into the build folder, a folder named tools is automatically created and contains useful tools. The list of tools currently available are the following:

• wcc: Computes, given a graph and a partition, the WCC of the partition.
  • Usage: wcc -f [graph file name] -p [partition file name]
• f1score: Computes the f1score between two partitions.
  • Usage: f1score [partition file name 1] [partition file name 2]