This is an offical implementation of FOIL: Time-Series Forecasting for Out-of-Distribution Generalization Using Invariant Learning.
Dependencies can be installed using the following file: newtimelib_environment.yml
You can obtain the well pre-processed datasets from [Google Drive] or [Baidu Drive], Then place the downloaded data in the folder./dataset
Usecase Run Raw Informer on ILI dataset with Pred_Len=4:
cd Informer-Raw
python ILI-Pred4.py Run Informer with FOIL on ILI dataset with Pred_Len=4:
cd Informer+FOIL
python ILI-Pred4-0.py
python ILI-Pred4-1.py - First Infer Envrionment; Second Learn Invariant Reperesentation
If you find this repo useful, please cite our paper.
@inproceedings{
liu2024timeseries,
title={Time-Series Forecasting for Out-of-Distribution Generalization Using Invariant Learning},
author={haoxin liu and Harshavardhan Kamarthi and Lingkai Kong and Zhiyuan Zhao and Chao Zhang and B. Aditya Prakash},
booktitle={Forty-first International Conference on Machine Learning},
year={2024},
url={https://openreview.net/forum?id=SMUXPVKUBg}
}
If you have any questions or suggestions, feel free to contact: hliu763@gatech.edu
This library is constructed based on the following repos:
https://github.com/zhouhaoyi/Informer2020/