SHIELD Dataset
收藏DataCite Commons2026-04-30 更新2026-05-05 收录
下载链接：
https://stanford.redivis.com/datasets/ee4n-6gv2f84et?v=1.0
下载链接
链接失效反馈官方服务：
资源简介：
IMPORTANT
- To start the process and verify your institutional affiliation, please create a redivis account using your institutional email.
- You must complete all the steps before we grant you access as a member
SHIELD is a dataset of clinical text with PHI (Protected Health Information) span annotations for benchmarking de-identification and named entity recognition models. The dataset contains de-identified clinical notes from Stanford Medicine with expert-annotated PHI spans across 9 entity categories: age, date, doctor, hospital, id, location, patient, phone, and web.
Access Requirements:
- Institutional email address (personal emails will be rejected)
- CITI Training: "Data or Specimens Only Research" course with all 13 modules (same as MIMIC-IV on PhysioNet). Instructions: https://physionet.org/about/citi-course/
- Signed Data Use Agreement (SHIELD Data Set License 1.0)
Tables:
- notes: note_id, note_text, note_type
- spans: span_id, note_id, span_start, span_end, span_label
License: SHIELD Data Set License 1.0 (modeled after PhysioNet v1.5.0)
Data must remain on encrypted machines. Redistribution is forbidden.
Contact: JD Posada (jdposada@stanford.edu)
提供机构：
Redivis
创建时间：
2026-04-02
