LFR-benchmark

The LFR benchmark model for networks. This repository contains a Python wrapper for the original code in C (original code). I'd like to thank the authors for generously allowing me to create this repository. If you use this code, please cite the following paper:

@article{PhysRevE.78.046110,
  title = {Benchmark graphs for testing community detection algorithms},
  author = {Lancichinetti, Andrea and Fortunato, Santo and Radicchi, Filippo},
  journal = {Phys. Rev. E},
  volume = {78},
  issue = {4},
  pages = {046110},
  numpages = {5},
  year = {2008},
  month = {Oct},
  publisher = {American Physical Society},
  doi = {10.1103/PhysRevE.78.046110},
  url = {https://link.aps.org/doi/10.1103/PhysRevE.78.046110}
}

How to install

To install the LFR-benchmark, follow these steps:

git clone https://github.com/skojaku/LFR-benchmark
cd LFR-benchmark
python setup.py build
pip install -e .

Usage

Once you have installed the LFR-benchmark, you can use it in your Python code as follows:

import lfr

params = {
    "N": 1000,     # number of nodes
    "k": 50,       # average degree
    "maxk": 100,   # maximum degree
    "minc": 20,    # minimum community size
    "maxc": 100,   # maximum community size
    "tau": 3.0,    # degree exponent
    "tau2": 2.0,   # community size exponent
    "mu": 0.5,     # mixing rate
}

ng = lfr.NetworkGenerator()
data = ng.generate(**params)

net = data["net"]                  # scipy.csr_sparse matrix
community_table = data["community_table"]  # pandas DataFrame
seed = data["seed"]                # Seed value

In the code snippet above, you can modify the params dictionary to adjust the parameters for network generation according to your needs. After generating the benchmark network using the ng.generate() method, you can access the generated network (net) as a scipy.csr_sparse matrix, the community information (community_table) as a pandas DataFrame, and the seed value (seed) used for network generation.