Multilayer Louvain

This repository contains the code for our proposed "Multilayer Louvain" algorithm which is able to detect communities in multilayer networks. Simultaneously, we also provide the code to generate synthetic multilayer networks with ground truth communities. This kind of networks can be used for evaluating the performance of the community detection algorithms on multilayer networks.

Setup -

• We require python 2.7 and a handful of other supporting libraries. To install the other dependencies, use: pip install -r requirements.txt

Multilayer Louvain algorithm

• How to run? -
  • Parameters - Update the following two paths in Multilayer-Louvain.py as required,
    • Pathtosave - directory path to save result file
    • networkpath - directory path from where to read network files
  • Command - python Multilayer-Louvain.py

Synthetic Network Generation

• How to run? -
  • Parameters - Update the following variables in generate_networks.py,
    • list_alpha - Set of 'alpha' values for which networks need to be generated
    • list_p - Set of 'p' values for which networks need to be generated
    • list_p1 - Set of 'p1' values for which networks need to be generated
    • list_p2 - Set of 'p2' values for which networks need to be generated
    • n - Number of nodes in each layer
    • k - Average degree of the individual network layers
    • maxk - Maximum degree of the individual network layers
    • mu - Probabilty of having edges across communities
  • Command - python2.7 generate_networks.py <folder_name1> <folder_name2>
  • Note: Prior to running the above command, the 'permission of execution' would be required to be given to the 'benchmark' executable which generates the LFR networks.
  • The above command generates 2-layer multilayer networks in the folder named 'folder_name1_Networks' with all possible combinations of the specified parameter values. 'folder_name2' would contain the LFR networks,
  • The file names of the generated network files would be in the following format - network_<alpha>_<p>_<mu>_<p1>_<p2>
  • The format of the generated network files is the following -

    • number_of_layers

      layer1 vertices

      number_of_layer1_edges

      layer1_edge1

      layer1_edge2

      ...

      layer2 vertices

      number_of_layer2_edges

      layer2_edge1

      layer2_edge2

      ...

      number_of_couplings

      coupling1_top_layer

      coupling1_bot_layer

      number_of_coupling1_edges

      coupling1_edge1

      coupling1_edge2

      ...

      coupling2_top_layer

      coupling2_bot_layer

      number_of_coupling2_edges

      coupling2_edge1

      coupling2_edge2

      ...

      number_of_communities

      community1_vertices

      community2_vertices

      ...