Networked Time Series Imputation via Position-aware Graph Enhanced Variational Autoencoders (KDD 2023 - Link - Arxiv)

Datasets

• The datasets being used in the paper can be found in this link.

• After downloading and unzipping the datasets, please move them into the dataset folder under the root of this repo.

Run Experiments

• run_imputation.py is used to compute the metrics for the deep imputation methods. An example of usage is

   python run_imputation.py --config config/pogevon/air36.yaml --in-sample False
  

• When running experiments for PEMS-BA, PEMS-LA and PEMS-SD datasets, one needs to change the subdataset_name value in config file pems.ymal to 'PEMS-04', 'PEMS-07' and 'PEMS-11' respectively.

Requirements

We run all the experiments in python 3.8, see requirements.txt for the list of pip dependencies.

Bibtex reference

If you find this code useful please consider to cite our paper:

@inproceedings{wang2023networked,
  title={Networked Time Series Imputation via Position-aware Graph Enhanced Variational Autoencoders},
  author={Wang, Dingsu and Yan, Yuchen and Qiu, Ruizhong and Zhu, Yada and Guan, Kaiyu and Margenot, Andrew and Tong, Hanghang},
  booktitle={Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining},
  pages={2256--2268},
  year={2023}
}

Acknowledgement

This repo is based on the implementations of GRIN and thanks for their contribution.