LAD

Official python implementation of the paper: Laplacian Change Point Detection for Dynamic Graphs (KDD 2020)

For more info on me and my work, please checkout my website.

If you have any questions, feel free to contact me at my email: [email protected]

Many thanks to my amazing co-authors: Yasmeen Hitti, Guillaume Rabusseau, Reihaneh Rabbany

Follow-up Works

• Scale to Large Dynamic Graphs: Fast and Attributed Change Detection on Dynamic Graphs (PAKDD 2023), github

• Extending to Multi-view Dynamic Graphs: Laplacian Change Point Detection for Single and Multi-view Dynamic Graphs (preprint), github

Content:

all synthetic experiments and real world experiments from the paper can be reproduced here.

Datasets:

In datasets/, You can find edgeslists for both the synthetic and real world experiments we have.

In datasets/canVote_processed, you can find our original Canadian Bill Voting network. if you use it, please cite this paper.

Usage:

1. first extract the edgelists in datasets/SBM_processed/hybrid, pure, resampled.zip

2. To reproduce synthetic experiments (-n is the number of eigenvalues used)

• python SBM_Command.py -f pure -n 499

substitute pure with hybrid or resampled for the corresponding settings

3. To reproduce real world experiments

• python Real_Command.py -d USLegis -n 6

• python Real_Command.py -d UCI -n 6

• python Real_Command.py -d canVote -n 338

Library:

1. python 3.8.1

2. scipy 1.4.1

3. scikit-learn 0.22.1

4. tensorly 0.4.5

5. networkx 2.4

6. matplotlib 1.3.1

Citation:

If code or data from this repo is useful for your project, please consider citing our paper:

@inproceedings{huang2020laplacian,
  title={Laplacian Change Point Detection for Dynamic Graphs},
  author={Huang, Shenyang and Hitti, Yasmeen and Rabusseau, Guillaume and Rabbany, Reihaneh},
  booktitle={Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery \& Data Mining},
  pages={349--358},
  year={2020}
}