zhangyashen / conad

Contrastive Attributed Network Anomaly Detection with Data Augmentation (PAKDD'22)

Get Started

This repository contains the code for our PAKDD'22 paper, Contrastive Attributed Network Anomaly Detection with Data Augmentation, as well as three other GNN baselines used as comparison. The evaluation data are included in the data/ folder. To run these models, use the main.py script and corresponding training function in it.

Download Data

• The datasets used in this paper are stored as git lfs objects, and can be downloaded with git lfs pull.
• In case the git-lfs storage quota runs out, download the datasets from Google Drive.

Cite

Please cite our paper if you use the model or this code in your own work:

@inproceedings{xu_conad_2022,
  title = {Contrastive Attributed Network Anomaly Detection with Data Augmentation},
  author = {Xu, Zhiming and Huang, Xiao and Zhao, Yue and Dong, Yushun and Li, Jundong},
  booktitle={Pacific-Asian Conference on Knowledge Discovery and Data Mining (PAKDD)},
  year = {2022}
}