Computing node centralities


Introduction: A node centrality measure computes the topological importance of a node in the network (or graph). Various centrality measures have been used to link topological importance of nodes in the network to its functional importance. With this motivation, we study topologies of nodes in networks using the following seven state-of-the-art node centrality measures in our dynamic network analysis paper.

  1. Degree centrality (D)

  2. Clustering coefficient centrality (CC)

  3. k-coreness centrality (K)

  4. Graphlet degree centrality (G)

  5. Betweenness centrality (B)

  6. Closeness centrality (C)

  7. Eccentricity centrality (EC)

Below we provide the implementation of the seven node centrality measures from our following paper.

Reference: Fazle E. Faisal and Tijana Milenkovic (2014), Dynamic networks reveal key players in aging, Bioinformatics, 30(12):1721-1729.

Software: The Unix version implementation for computing node centralities is available here.

Usage: ./ [input_network] [input_network_format] [node_centrality_measure] [output_directory]

  • [input_network_format] can be either "E" (without quotes) for edge list format or "L" (without quotes) for LEDA format. The details on LEDA format are available here.

  • [node_centrality_measure] can be "D", "CC", "K", "G", "B", "C", or "EC" (without quotes). See above for the list of seven node centrality measures. All of the seven node centralities can be computed at once with option "A" (without quotes).

Example: ./ sample_network/sample.txt E A sample_output

  • The command will compute all of the seven node centralities for the input file "sample_network/sample.txt" and will produce seven output files named sample.degc, sample.clusc, sample.kc, sample.gdc, sample.betwc, sample.closec, and sample.ecc within the output directory "sample_output".

  • Note. The sample input and output files are provided in our software.

Please cite our our paper if you use the software.