Scribd Logo
Upload

Welcome to Scribd!

  • 5 views

    MEMELink

    Uploaded by

    ariyan.hossain2208
    The document provides instructions and links to resources for running experiments on community detection algorithms using synthetic and real-world network datasets. It includes links to download benchmark network generators, community detection code, evaluation metrics, and sample network datasets. Instructions are provided on running the MEMELink community detection algorithm and evaluating the results using normalized mutual information.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    MEMELink

    Uploaded by

    ariyan.hossain2208
    5 views3 pages
    The document provides instructions and links to resources for running experiments on community detection algorithms using synthetic and real-world network datasets. It includes links to download benchmark network generators, community detection code, evaluation metrics, and sample network datasets. Instructions are provided on running the MEMELink community detection algorithm and evaluating the results using normalized mutual information.

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    The document provides instructions and links to resources for running experiments on community detection algorithms using synthetic and real-world network datasets. It includes links to download benchmark network generators, community detection code, evaluation metrics, and sample network datasets. Instructions are provided on running the MEMELink community detection algorithm and evaluating the results using normalized mutual information.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    5 views3 pages

    MEMELink

    Uploaded by

    ariyan.hossain2208
    The document provides instructions and links to resources for running experiments on community detection algorithms using synthetic and real-world network datasets. It includes links to download benchmark network generators, community detection code, evaluation metrics, and sample network datasets. Instructions are provided on running the MEMELink community detection algorithm and evaluating the results using normalized mutual information.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 3

    Please, read this first

    I won’t be able to assist you with coding, debugging, installation, setup or experiments. The provided files
    are "as is." I've done my best to include all the info you'll need. I trust that the Thesis and this document will
    guide you well for your experiments. Best of luck!

    The pseudocode to write your algorithms are available here:


    https://nova.newcastle.edu.au/vital/access/manager/Repository?view=list&f0=sm_creator%3A%22Gabardo%2
    C+Ademir+Cristiano%22&sort=sort_ss_sm_creator%5C

    Important related code sources to perform the experiments:

    1. The LFR synthetically generated benchmark networks (section 3.3.3 of my PhD Thesis), you will find
    the Generators at the section (LFR benchmark graphs) here: https://www.santofortunato.net/resources

    Some extra help about the parameters here:


    https://networkx.org/documentation/stable/reference/generated/networkx.generators.community.LFR_benchma
    rk_graph.html

    2. The NMI (normalised mutual information), the onmi executable file in the folder, comes from here:
    https://github.com/aaronmcdaid/Overlapping-NMI (usage examples are there as well as per publications about
    the NMI), you can download and compile the executable with GCC.

    3. The source for the “organic” datasets are (although I left some in the folder as well, you may want them
    in a different format):
    a. http://www-personal.umich.edu/~mejn/netdata
    b. https://deim.urv.cat/~alexandre.arenas/data/welcome.htm

    Instructions about how to run MEMELink

    1. Make sure that you have Java properly installed in your computer

    2. Extract the contents of the .zip file in a folder of your preference. It should looks like this, the
    MEMELINK.jar goes along a MEMElink folder with datasets, parameters, etc.
    3. Run the .jar file with the following command.

    java -jar MEMELINK.jar

    You should see a similar output as:

    The visualisation tool used is the GraphStream Java Library, you can find it here:
    https://github.com/graphstream

    As you run the algorithm, the output will be saved in the dataset folder, for instance. If in the parameters you
    chose to run 5 times,
    file:MEMELink/datasets/karate/karate.net
    runs:5
    verbose:false
    display:true
    stats:true
    popSize:100
    maxGens:100
    seedtype:4

    There will be 10 output files in the dataset directory, two for each run, the main one is without any post
    processing, and the _post files are with post processing.

    The format of these output files is a community per line. For instance, in this output for the karate network:

    2 1 3 4 5 6 7 8 9 10 11 12 13 14 17 18 20 22 28 29 31 32 33 15 16 21 23 34
    26 24 25 28 32
    30 24 27 33 34
    33 19 34

    There are four communities, each entry in the line represents the nodes in that community. In this example the
    community in the fourth line has nodes 33 19 34. These are the files used to compare the results using the
    NMI. Additionally, there will be a fitness.txt file with the fitness for each run.

    For the Xie synthetic networks, use the GT.txt (ground truth) files to compute the NMI against the generated
    MML files, eg: ../../onmi GT.txt MML_1.txt There is an automate_onmi.sh that makes running in
    batches easier.

    You might also like