Correlation Clustering Algorithm for Dynamic Complete Signed Graphs: An Index-based Approach

Published version at Knowledge and Information Systems:

@article{shakiba2025correlation,
  title={Correlation clustering algorithm for dynamic complete signed graphs: an index-based approach},
  author={Shakiba, Ali},
  journal={Knowledge and Information Systems},
  pages={1--20},
  year={2025},
  publisher={Springer London London},
  doi={10.1007/s10115-025-02397-3}
}

Source code for the following paper:

@misc{https://doi.org/10.48550/arxiv.2301.00384,
  doi = {10.48550/ARXIV.2301.00384},
  url = {https://arxiv.org/abs/2301.00384},
  author = {Shakiba, Ali},
  keywords = {Data Structures and Algorithms (cs.DS), Machine Learning (cs.LG), FOS: Computer and information sciences, FOS: Computer and information sciences},
  title = {Correlation Clustering Algorithm for Dynamic Complete Signed Graphs: An Index-based Approach},
  publisher = {arXiv},
  year = {2023}
}

Requirements

• boost library: the boost_log library for logging, tested with version 1.74 shipped with ubuntu (sudo apt install sudo apt install libboost-log1.74.0 libboost-log1.74-dev)
• bazel build system
• Google Test Suit

Running the experiments

• A sample running workflow:

  1. Run the preprocess.py for the datasets downloaded from the SNAP
  2. Then, run the following command to do a all-in-one experiment for a dataset: bazel run //main:all input_filename output_prefix default_eps batch-auto eps-schedule-length.

• There are two general forms for running this code in batch mode:

  1. Using a pre-generated epsilon schedule in a file: bazel run //main:all [input_filename] [output_prefix] [default_eps] [batch] [eps-schedule-file].
  2. Using the distribution of the non-agreement among the edges to generate a equal-width epsilon schedule: bazel run //main:all [input_filename] [output_prefix] [default_eps] [auto-batch] [eps-schedule-length].

• If you want to run the experiments interactively, then just run ```bazel run //main:all [input_filename] [output_prefix] [default_eps]``.

• All the output files would be put in the data\*.out files tagged with the output_prefix.

Documentation

Use doxygen to generate the docmentation for the project.

doxygen Doxyfile

Testing

For testing, just run the following set of commands

bazel test --test_output=all --show_progress  //tests:graph_test