A dataset for Semantic Textual Similarity (STS) in Japanese Clinical Domain Texts

This dataset contains of sentence pairs extracted from CiNii (https://ci.nii.ac.jp/).

Each sentence pair is annotated with a similarity score 0 (low semantic similarity) to 5 (high semantic similarity).

Citation

If you use this dataset please cite our paper:

@article{mutinda2021semantic,
  title={Semantic Textual Similarity in Japanese Clinical Domain Texts Using BERT},
  author={Mutinda, Faith Wavinya and Yada, Shuntaro and Wakamiya, Shoko and Aramaki, Eiji},
  journal={Methods of Information in Medicine},
  year={2021},
  publisher={Georg Thieme Verlag KG}
  }