In this article, researchers aimed to develop high-quality annotations for NLP by selecting clinical notes from diverse patients across two institutions. They focused on enhancing the annotation corpus with terms related to Social Support (SS) and Social Isolation (SI) by selecting notes that contained specific lexicon terms. By leveraging a variety of note-writers and contexts, the aim was to avoid redundancy and improve the robustness of the NLP pipelines. The annotation process involved the use of the Brat Rapid Annotation Tool, ensuring consistency across different note settings.
To create gold standard data for developing NLP pipelines, we selected notes from diverse unique patients to enrich annotation for Social Support and Social Isolation.
Notes were chosen to maximize contextual diversity by including different note-writers and avoiding redundancy. This strategy aimed to provide comprehensive data for improved NLP applications.
Collection
[
|
...
]