/webMedQA

A Chinese medical question answering dataset

Apache License 2.0Apache-2.0

webMedQA

A real-world Chinese medical question answering dataset collected from online health consultancy websites. Our paper

Dataset description

Train Dev Test
Questions 50610 6337 6337
Avg length 86.68 87.43 86.08
Answers 253050 21685 31685
Avg length 146.88 147.74 148.50

Each question has 1 positive and 4 negative answers. A sample:

sample

Please read our paper for more detail.

Please Cite

@article{he2019applying,
  title={Applying deep matching networks to Chinese medical question answering: A study and a dataset},
  author={He, Junqing and Fu, Mingming and Tu, Manshu},
  journal={BMC Medical Informatics and Decision Making},
  volume={19},
  number={2},
  pages={52},
  year={2019},
  doi={10.1186/s12911-019-0761-8}
}