Multivariate Time Series Anomaly Detection from the Perspective of Graph Relational Learning
This repository contains the implementation of
Zhang, W., Zhang, C., and Tsung, F. " GRELEN: Multivariate Time Series Anomaly Detection from the Perspective of Graph Relational Learning ," Proceedings of the 31st International Joint Conference on Artificial Intelligence and the 25th European Conference on Artificial Intelligence (IJCAI-ECAI 2022) [PDF]
Four datasets are included in our paper:
SWAT and WADI: [[download-from-iTrust]](iTrust Labs_Dataset Info - iTrust)
SMD: [download-from-OmniAnomaly]
PSM: [[download-from-RANSynCoders]](GitHub - eBay/RANSynCoders)
- For training:
python train_grelen.py
- For testing:
python test_grelen.py
The config files for datasets, model and evaluation should be put in the config_files folder. Config files should contain the data path, model hyper-parameters, training settings for model training. Target model (parameter path) should be included while model testing.