SDOH-NLI
收藏arXiv2023-10-28 更新2024-06-21 收录
下载链接：
https://github.com/google-research-datasets/SDOH-NLI
下载链接
链接失效反馈官方服务：
资源简介：
SDOH-NLI数据集由谷歌研究创建，旨在从临床笔记中提取社会健康决定因素（SDOH）。该数据集包含29,635条标注实例，基于公开的医疗报告，通过自然语言推理（NLI）任务形式化，使用人工标注的二元文本蕴含标签。数据集的创建过程涉及从MTSamples.com抓取5,003份医疗报告，筛选出1,398个社会历史片段，并手动整理60个SDOH因素。SDOH-NLI数据集的应用领域包括帮助医疗提供者系统地识别改善健康结果的机会，并解决健康差异问题。
The SDOH-NLI dataset was developed by Google Research to extract social determinants of health (SDOH) from clinical notes. It comprises 29,635 annotated instances, which are formalized as natural language inference (NLI) tasks based on publicly available medical reports, with manually annotated binary textual entailment labels. The dataset creation workflow involved scraping 5,003 medical reports from MTSamples.com, screening 1,398 social history snippets, and manually curating 60 SDOH factors. Application scenarios of the SDOH-NLI dataset include assisting healthcare providers in systematically identifying opportunities to improve health outcomes and addressing health disparities.
提供机构：
谷歌研究
创建时间：
2023-10-28
