Neural additive time-series models: Explainable deep learning for multivariate time-series prediction (NATMs)

This repo is the Pytorch implementation of NATMs.

How to use?

We provide the jupyter notebook file (.ipynb) for learning and inference process.
Also, python script registered but we recommend to use the notebook.

Performance

Citation

@article{JO2023120307,
title = {Neural additive time-series models: Explainable deep learning for multivariate time-series prediction},
journal = {Expert Systems with Applications},
volume = {228},
pages = {120307},
year = {2023},
issn = {0957-4174},
doi = {https://doi.org/10.1016/j.eswa.2023.120307},
author = {Wonkeun Jo and Dongil Kim},
keywords = {Deep learning, Explainable artificial intelligence,Multivariate time series prediction, Neural additive models, Parameter sharing},
}