---
license: cc-by-nc-4.0
task_categories:
- text-classification
- token-classification
language:
- ja
---


USA model paper SCPOS datasets. Including four sub-datasets.

paper address:
https://arxiv.org/abs/2309.03787

Cite:

@article{gan2023usa,
  title={USA: Universal Sentiment Analysis Model \& Construction of Japanese Sentiment Text Classification and Part of Speech Dataset},
  author={Gan, Chengguang and Zhang, Qinghao and Mori, Tatsunori},
  journal={arXiv preprint arXiv:2309.03787},
  year={2023}
}

This dataset constructed base in JGLUE benchmark text sentiment classification task
https://github.com/yahoojapan/JGLUE