yuanhetian/chimed

ChiMed

This github repository includes the information about the ChiMed Corpus.

🔥 Updates 🔥

Nov. 19, 2023

We recently released a large language model for the Chinese medical domain named ChiMed-GPT, which is trained on ChiMed data. For more information, please visit our GitHub Repo.

Apr. 29, 2022

ChiMed is availabel by valid request! The ChiMST corpus contains 1,000 QA pages with annotations of Chinese word segmentation and medical terms.

The Copyright

The copyright of the corpus belongs to 39ask. We release the ChiMed corpus based on our contract with 39ask.

Request the ChiMed and ChiMST Dataset

Please vist here for the information to request the datasets.

Important Things to Note before Using the ChiMed

The Statistics

	Number
# of QA pages	200,744
# of departments	15
# of questions	200,744
avg # of characters per question	55.57
# of answers	401,488
avg # of characters per answer	85.21
# of unique keyphrases	11,724
avg # of keyphrases per QA page	4.51

"Recommended" flag vs. "Adopted" flag

For each answer in the corpus, there are two flags: Recommended and Adopted. Their differences are:

• Recommended: whether the answer is recommended by the 39ask website (chosen by the website system);
• Adopted: whether the patient adopts the answer (chosen by the user).

Citation

If you use the ChiMed corpus, please cite the following paper (Note: the ChiMed Corpus is larger than the dataset used in this paper).

@inproceedings{tian-etal-2019-chimed,
    title = "ChiMed: A Chinese Medical Corpus for Question Answering",
    author = "Tian, Yuanhe and Ma, Weicheng and Xia, Fei and Song, Yan",
    booktitle = "Proceedings of the 18th BioNLP Workshop and Shared Task",
    month = aug,
    year = "2019",
    address = "Florence, Italy",
    pages = "250--260",
}