SimCMR is a network community detection algorithm particularly developed for detecting communities in large-scale networks fast and effectively. It was designed for detecting non-overlapping communities in unipartite, undirected, and unweighted networks.
SimCMR was implemented in pure Python on top of NetworkX which is a very popular network analysis framework written purely in Python. You can find my own implementation in code directory (tunali_community.py). In the same folder, you can find the Python code that I have used to generate LFR benchmark networks as well as the code to run the experiments explained in the published research paper.
In data directory, I share all real-world and artificial benchmark networks in a unified format (edge list with node indices starting from 1). In addition to network files, you can find their corresponding ground-truth community assignments.
In executables directory, I share all executables that I have used to generate LFR benchmarks and some implementations of other community detection algorithms I have compared with SimCMR in my research.
If you find any material here useful, please cite my research as below:
V. Tunali, "Large-Scale Network Community Detection Using Similarity-Guided Merge and Refinement," IEEE Access, vol. 9, pp. 78538-78552, 2021, doi: 10.1109/ACCESS.2021.3083971.
Link to the paper on IEEE Xplore