The link of paper : https://arxiv.org/pdf/2005.04809.pdf
https://github.com/datasets/covid-19/tree/master/data
Please modify the data pre-processing by yourself, for example if the duration of each sample is more than 100 days, you should modify the "split_sequences" function inside the code such that the proportion of input and label is around 75:25 or 70:30
@misc{yudistira2020covid19,
title={COVID-19 growth prediction using multivariate long short term memory},
author={Novanto Yudistira},
year={2020},
eprint={2005.04809},
archivePrefix={arXiv},
primaryClass={cs.LG}
}