Official implementation of SiGMA.

TL;DR

• We propose SiGMA, a principled multi-scale time series forecasting framework that overcomes the limitations of existing multi-scale forecasting approaches.
• This project is developed based on experiments using TSLib, which is distributed under the MIT License.

Environment Setup

conda create -n SiGMA python=3.8.0
conda activate SiGMA
pip install -r requirements.txt

How to Run?

bash script/SiGMA_[dataset_name].sh

Supported dataset_name values: ECL, ETTh1, ETTh2, ETTm1, ETTm2, Exchange, M4, Traffic, Weather

You can download these datasets from TSLib.

License

This project is licensed under the MIT License - see the LICENSE file for details.