Obstetrics Notes Collection (ONC)
收藏资源简介:
ONC是由伊利诺伊大学芝加哥分校团队构建的产科临床笔记数据集,包含100份经过脱敏处理的病史与体检记录,涵盖阴道分娩和重复剖宫产患者。该数据集通过REDCap系统收集,采用Spark NLP框架自动脱敏,并经由医学专家人工标注章节边界,保留了产科特有的'妊娠史'等专业章节标题。作为对MIMIC-III公共语料库的领域补充,ONC专注于评估模型在产科这一临床子领域的跨域适应能力,为医疗NLP在稀缺标注数据场景下的应用提供基准测试资源。
ONC is an obstetric clinical note dataset constructed by the team from the University of Illinois Chicago. It contains 100 de-identified medical history and physical examination records, covering patients who underwent vaginal delivery and repeat cesarean section. This dataset was collected via the REDCap system, automatically de-identified using the Spark NLP framework, with section boundaries manually annotated by medical experts, and retains professional section titles unique to obstetrics such as "Pregnancy History". As a domain-specific supplement to the MIMIC-III public corpus, ONC focuses on evaluating the cross-domain adaptation capability of models in the clinical subfield of obstetrics, providing benchmark resources for the application of medical NLP in scenarios with scarce annotated data.
- 1Bridging the Domain Divide: Supervised vs. Zero-Shot Clinical Section Segmentation from MIMIC-III to Obstetrics伊利诺伊大学芝加哥分校 · 2026年



