De-identification of psychiatric intake records: Overview of 2016 CEGS N-GRID shared tasks Track 1.

Amber Stubbs,Michele Filannino,Özlem Uzuner

Journal of Biomedical Informatics（2017）

引用 120|浏览459

暂无评分

摘要

•NLP shared task with new set of 1000 de-identified psychiatric records.•“Sight-unseen” task: top F1 of 0.799 using out-of-the-box system on new data.•“Standard task: top F1 of 0.914 on test data after 2months of development.•Hybrid systems most effective, but often missed PHI requiring world knowledge or context.

查看译文

关键词

Natural language processing,Machine learning,Clinical records,Shared task

AI 理解论文

溯源树

样例

生成溯源树，研究论文发展脉络

Chat Paper

正在生成论文摘要